Now look at this! (a bit of a rant at FF and page designers)
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 top - 02:40:02 up 11 days, 17:10, 2 users, load average: 1,47, 1,61, 1,52 Tasks: 652 total, 2 running, 649 sleeping, 0 stopped, 1 zombie %Cpu(s): 11,2 us, 0,9 sy, 0,6 ni, 86,5 id, 0,2 wa, 0,0 hi, 0,6 si, 0,0 st KiB Mem : 32821800 total, 2750532 free, 28228940 used, 1842328 buff/cache KiB Swap: 10485760+total, 95023808 free, 9833788 used. 3340160 avail Mem PID USER PR NI VIRT RES SHR SWAP S %CPU %MEM TIME+ COMMAND 11012 cer 20 0 14,010g 0,010t 81556 167372 R 100,9 33,92 670:39.60 Web Content <======== 11096 cer 20 0 9982108 4,464g 119580 993160 S 2,985 14,26 306:24.60 Web Content 32435 cer 20 0 8121244 2,753g 92956 915588 S 0,000 8,796 302:12.75 thunderbird-bin 10799 cer 20 0 6830940 1,331g 506812 454940 S 1,791 4,251 401:56.06 firefox 11152 cer 20 0 5157264 829856 71600 175304 S 1,194 2,528 102:52.05 Web Content 10942 cer 20 0 4462244 704732 58320 176224 S 0,896 2,147 119:38.94 Web Content 11068 cer 20 0 5306312 644124 100104 534320 S 0,896 1,962 124:44.93 Web Content 10990 cer 20 0 4281224 586028 111408 154048 S 6,866 1,785 210:21.72 Web Content 4206 cer 20 0 4019828 564988 15012 0 S 0,299 1,721 6:00.57 shotwell See that process using 0,010t of resident memory? I've never seen the word "teras" of ram used for a single process... :-o It was probably this page: <https://www.movistar.es/particulares/oferta-combinada/fusion> just a list of packages from my ISP... I have another tab that Firefox about:performance says it is using 2.33GB of ram. A train information page: <https://www.thetrainline.com/en/train-times/seville-to-berlin> what a lot of crap, how can they design those pages that eat so many resources. - -- Cheers Carlos E. R. (from 15.2 x86_64 at Telcontar) -----BEGIN PGP SIGNATURE----- iHkEARECADoWIQQZEb51mJKK1KpcU/W1MxgcbY1H1QUCYFqbahwccm9iaW4ubGlz dGFzQHRlbGVmb25pY2EubmV0AAoJELUzGBxtjUfVxGMAn3VfxZV19w7GaTc1CW2h OV8qDQZ9AJiRKzg2WVHSRODK9mm6rCmW0q2z =sPPH -----END PGP SIGNATURE-----
On 2021-03-23 20:52:42 Carlos E. R. wrote:
|top - 02:40:02 up 11 days, 17:10, 2 users, load average: 1,47, 1,61, | 1,52 Tasks: 652 total, 2 running, 649 sleeping, 0 stopped, 1 zombie | %Cpu(s): 11,2 us, 0,9 sy, 0,6 ni, 86,5 id, 0,2 wa, 0,0 hi, 0,6 si, | 0,0 st KiB Mem : 32821800 total, 2750532 free, 28228940 used, 1842328 | buff/cache KiB Swap: 10485760+total, 95023808 free, 9833788 used. | 3340160 avail Mem | | PID USER PR NI VIRT RES SHR SWAP S %CPU %MEM | TIME+ COMMAND 11012 cer 20 0 14,010g 0,010t 81556 167372 R 100,9 | 33,92 670:39.60 Web Content <======== 11096 cer 20 0 9982108 | 4,464g 119580 993160 S 2,985 14,26 306:24.60 Web Content 32435 cer | 20 0 8121244 2,753g 92956 915588 S 0,000 8,796 302:12.75 | thunderbird-bin 10799 cer 20 0 6830940 1,331g 506812 454940 S | 1,791 4,251 401:56.06 firefox 11152 cer 20 0 5157264 829856 | 71600 175304 S 1,194 2,528 102:52.05 Web Content 10942 cer 20 0 | 4462244 704732 58320 176224 S 0,896 2,147 119:38.94 Web Content 11068 | cer 20 0 5306312 644124 100104 534320 S 0,896 1,962 124:44.93 Web | Content 10990 cer 20 0 4281224 586028 111408 154048 S 6,866 1,785 | 210:21.72 Web Content 4206 cer 20 0 4019828 564988 15012 0 | S 0,299 1,721 6:00.57 shotwell | | | | |See that process using 0,010t of resident memory? I've never seen the word |"teras" of ram used for a single process... :-o | |It was probably this page: | |<https://www.movistar.es/particulares/oferta-combinada/fusion> | |just a list of packages from my ISP... | | |I have another tab that Firefox about:performance says it is using 2.33GB |of ram. A train information page: | |<https://www.thetrainline.com/en/train-times/seville-to-berlin> | | |what a lot of crap, how can they design those pages that eat so many |resources.
That's what we get from web developers who use WYSIWYG code generators and gratuitously include large volumes of javascript in their pages. Also, from what I've seen, every time there's an improvement in network bandwidth availability (POTS => ADSL => Cable => Fibre Optic), the size and complexity of web pages ballons another order of magnitude. It makes things especially frustrating for those of us who don't live in or near a major city, but in the country where the bandwidth improvements are years behind the cities'. Leslie -- openSUSE Leap 15.2 x86_64
On 2021/03/23 18:52, Carlos E. R. wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
KiB Mem : 32821800 total, 2750532 free, 28228940 used, 1842328 buff/cache KiB Swap: 10485760+total, 95023808 free, 9833788 used. 3340160 avail Mem
PID USER PR NI VIRT RES SHR SWAP S %CPU %MEM TIME+ COMMAND 11012 cer 20 0 14,010g 0,010t 81556 167372 R 100,9 33,92 670:39.60 Web Content <========
See that process using 0,010t of resident memory? I've never seen the word "teras" of ram used for a single process... :-o
10t of Resident memory used on a 32G machine? Um...
On 24/03/2021 05.07, L A Walsh wrote:
On 2021/03/23 18:52, Carlos E. R. wrote:
KiB Mem : 32821800 total, 2750532 free, 28228940 used, 1842328 buff/cache KiB Swap: 10485760+total, 95023808 free, 9833788 used. 3340160 avail Mem
PID USER PR NI VIRT RES SHR SWAP S %CPU %MEM TIME+ COMMAND 11012 cer 20 0 14,010g 0,010t 81556 167372 R 100,9 33,92 670:39.60 Web Content <========
See that process using 0,010t of resident memory? I've never seen the word "teras" of ram used for a single process... :-o
--- 10t of Resident memory used on a 32G machine? Um...
No, 0.01 teras of resident memory. In Spain, the ',' is the decimal separator. 14 G of virtual memory for a process. Just after sending the email, I had a tab that was not responding, and which was increasing half a gig per minute, of resident memory. I had to kill it that one and restart it. An openSUSE page: <https://connect.opensuse.org//pg/members/all/> -- Cheers / Saludos, Carlos E. R. (from 15.2 x86_64 at Telcontar)
Linda, et al -- ...and then L A Walsh said... % % On 2021/03/23 18:52, Carlos E. R. wrote: ... % > % >See that process using 0,010t of resident memory? I've never seen % >the word "teras" of ram used for a single process... :-o % --- % 10t of Resident memory used on a 32G machine? Um... Remember European vs American decimals and commas vs periods. That's 1/100 of 1T, or 10G, if I read it correctly. HTH & HAND :-D -- David T-G See http://justpickone.org/davidtg/email/ See http://justpickone.org/davidtg/tofu.txt
Carlos E. R. wrote:
See that process using 0,010t of resident memory? I've never seen the word "teras" of ram used for a single process... :-o
I have occasionally seen some crazy numbers in top, no relation to reality.
I have another tab that Firefox about:performance says it is using 2.33GB of ram. A train information page:
<https://www.thetrainline.com/en/train-times/seville-to-berlin>
I have just loaded it, 28Mb. -- Per Jessen, Zürich (11.1°C)
Per Jessen wrote:
Carlos E. R. wrote:
See that process using 0,010t of resident memory? I've never seen the word "teras" of ram used for a single process... :-o
I have occasionally seen some crazy numbers in top, no relation to reality.
I have another tab that Firefox about:performance says it is using 2.33GB of ram. A train information page:
<https://www.thetrainline.com/en/train-times/seville-to-berlin>
I have just loaded it, 28Mb.
And now, approx 90mins later, it has grown to 39Mb. -- Per Jessen, Zürich (13.2°C) http://www.dns24.ch/ - free dynamic DNS, made in Switzerland.
Per Jessen wrote:
Per Jessen wrote:
Carlos E. R. wrote:
See that process using 0,010t of resident memory? I've never seen the word "teras" of ram used for a single process... :-o
I have occasionally seen some crazy numbers in top, no relation to reality.
I have another tab that Firefox about:performance says it is using 2.33GB of ram. A train information page:
<https://www.thetrainline.com/en/train-times/seville-to-berlin>
I have just loaded it, 28Mb.
And now, approx 90mins later, it has grown to 39Mb.
approx 19:00, 100Mb+. I don't think the pages "are designed to eat so many resources", it looks more like a memory leak to me. javascript? -- Per Jessen, Zürich (9.5°C) http://www.dns24.ch/ - your free DNS host, made in Switzerland.
On 24/03/2021 19.05, Per Jessen wrote:
Per Jessen wrote:
Per Jessen wrote:
Carlos E. R. wrote:
See that process using 0,010t of resident memory? I've never seen the word "teras" of ram used for a single process... :-o
I have occasionally seen some crazy numbers in top, no relation to reality.
I have another tab that Firefox about:performance says it is using 2.33GB of ram. A train information page:
<https://www.thetrainline.com/en/train-times/seville-to-berlin>
I have just loaded it, 28Mb.
And now, approx 90mins later, it has grown to 39Mb.
approx 19:00, 100Mb+.
I don't think the pages "are designed to eat so many resources", it looks more like a memory leak to me. javascript?
Aha, so it grows. Mine was sitting for days, so it grew a lot. Yes, most pages use scripting of some sort nowdays. I don't know exactly how to find out. -- Cheers / Saludos, Carlos E. R. (from 15.2 x86_64 at Telcontar)
On Wed, 24 Mar 2021 19:10:02 +0100 "Carlos E. R." <robin.listas@telefonica.net> wrote:
On 24/03/2021 19.05, Per Jessen wrote:
Per Jessen wrote:
Per Jessen wrote:
Carlos E. R. wrote:
See that process using 0,010t of resident memory? I've never seen the word "teras" of ram used for a single process... :-o
I have occasionally seen some crazy numbers in top, no relation to reality.
I have another tab that Firefox about:performance says it is using 2.33GB of ram. A train information page:
<https://www.thetrainline.com/en/train-times/seville-to-berlin>
I have just loaded it, 28Mb.
And now, approx 90mins later, it has grown to 39Mb.
approx 19:00, 100Mb+.
I don't think the pages "are designed to eat so many resources", it looks more like a memory leak to me. javascript?
Aha, so it grows. Mine was sitting for days, so it grew a lot. Yes, most pages use scripting of some sort nowdays. I don't know exactly how to find out.
Using FF I don't see any growth so far, either with no JS or just with trainline enabled. Which domains do you have enabled? Use Tools/Web Developer/Toggle Tools/Memory
On 24/03/2021 21.24, Dave Howorth wrote:
On Wed, 24 Mar 2021 19:10:02 +0100 "Carlos E. R." <> wrote:
On 24/03/2021 19.05, Per Jessen wrote:
Per Jessen wrote:
Per Jessen wrote:
Carlos E. R. wrote:
See that process using 0,010t of resident memory? I've never seen the word "teras" of ram used for a single process... :-o
I have occasionally seen some crazy numbers in top, no relation to reality.
I have another tab that Firefox about:performance says it is using 2.33GB of ram. A train information page:
<https://www.thetrainline.com/en/train-times/seville-to-berlin>
I have just loaded it, 28Mb.
And now, approx 90mins later, it has grown to 39Mb.
approx 19:00, 100Mb+.
I don't think the pages "are designed to eat so many resources", it looks more like a memory leak to me. javascript?
Aha, so it grows. Mine was sitting for days, so it grew a lot. Yes, most pages use scripting of some sort nowdays. I don't know exactly how to find out.
Using FF I don't see any growth so far, either with no JS or just with trainline enabled. Which domains do you have enabled?
enabled? I don't understand. I don't block any domains, I don't block scripts. I only use uBlock.
Use Tools/Web Developer/Toggle Tools/Memory
-- Cheers / Saludos, Carlos E. R. (from 15.2 x86_64 at Telcontar)
Carlos E. R. wrote:
On 24/03/2021 19.05, Per Jessen wrote:
Per Jessen wrote:
Per Jessen wrote:
Carlos E. R. wrote:
See that process using 0,010t of resident memory? I've never seen the word "teras" of ram used for a single process... :-o
I have occasionally seen some crazy numbers in top, no relation to reality.
I have another tab that Firefox about:performance says it is using 2.33GB of ram. A train information page:
<https://www.thetrainline.com/en/train-times/seville-to-berlin>
I have just loaded it, 28Mb.
And now, approx 90mins later, it has grown to 39Mb.
approx 19:00, 100Mb+.
I don't think the pages "are designed to eat so many resources", it looks more like a memory leak to me. javascript?
Aha, so it grows. Mine was sitting for days, so it grew a lot.
The same page, now 258Mb. It must be a memory leak.
Yes, most pages use scripting of some sort nowdays. I don't know exactly how to find out.
If you want to know, just look at the HTML code - look for <script> tags, and/or meta links to .js files or type="text/javascript". Nice useful thing, that about:performance, I didn't know that one. Not sure if the "Energy Impact" column is very useful, maybe "elapse time" would be better :-) -- Per Jessen, Zürich (6.2°C)
On 25/03/2021 08.23, Per Jessen wrote:
Carlos E. R. wrote:
On 24/03/2021 19.05, Per Jessen wrote:
Per Jessen wrote:
Per Jessen wrote:
Carlos E. R. wrote:
See that process using 0,010t of resident memory? I've never seen the word "teras" of ram used for a single process... :-o
I have occasionally seen some crazy numbers in top, no relation to reality.
I have another tab that Firefox about:performance says it is using 2.33GB of ram. A train information page:
<https://www.thetrainline.com/en/train-times/seville-to-berlin>
I have just loaded it, 28Mb.
And now, approx 90mins later, it has grown to 39Mb.
approx 19:00, 100Mb+.
I don't think the pages "are designed to eat so many resources", it looks more like a memory leak to me. javascript?
Aha, so it grows. Mine was sitting for days, so it grew a lot.
The same page, now 258Mb. It must be a memory leak.
Yes, most pages use scripting of some sort nowdays. I don't know exactly how to find out.
If you want to know, just look at the HTML code - look for <script> tags, and/or meta links to .js files or type="text/javascript".
Ah. And I see also "<script>" For example, it contains this right at the start: «<script> if("PerformancePaintTiming" in window){var observer=new PerformanceObserver(function(r){var e=!0,n=!1,t=void 0;try{for(var a,i=r.getEntries()[Symbol.iterator]();!(e=(a=i.next()).done);e=!0){var o=a.value;"first-contentful-paint"===o.name&&(window.fcp=Math.round(o.startTime+o.duration))}}catch(r){n=!0,t=r}finally{try{e||null==i.return||i.return()}finally{if(n)throw t}}});observer.observe({entryTypes:["paint"]})} !function(){if("PerformanceLongTaskTiming" in window){var g=window.__tti={e:[]}; g.o=new PerformanceObserver(function(l){g.e=g.e.concat(l.getEntries())}); g.o.observe({entryTypes:["longtask"]})}}(); </script>» or this: «<script src="https://cdn.speedcurve.com/js/lux.js?id=401238997" async defer crossorigin="anonymous"></script>»
Nice useful thing, that about:performance, I didn't know that one. Not sure if the "Energy Impact" column is very useful, maybe "elapse time" would be better :-)
"energy" may be interesting for devices on battery, either laptops or phones. I take it as rough equivalent to CPU. Sorting by the memory indicator shows amazing things, like pages eating a gigabyte. -- Cheers / Saludos, Carlos E. R. (from 15.2 x86_64 at Telcontar)
Per Jessen wrote:
Carlos E. R. wrote:
See that process using 0,010t of resident memory? I've never seen the word "teras" of ram used for a single process... :-o
I have occasionally seen some crazy numbers in top, no relation to reality.
I have another tab that Firefox about:performance says it is using 2.33GB of ram. A train information page:
<https://www.thetrainline.com/en/train-times/seville-to-berlin>
I have just loaded it, 28Mb.
And now, approx 90mins later, it has grown to 39Mb.
Maybe some ads playing nasty? I have a windows machine here that sometimes crashes the ghostery plugin in the browser. Have to kill the chrome process and reload the plugin.
-- Per Jessen, Zürich (13.2°C) http://www.dns24.ch/ - free dynamic DNS, made in Switzerland.
-- L. de Braal BraHa Systems NL - Terneuzen T +31 115 649333
On 24/03/2021 20.30, Leen de Braal wrote:
Per Jessen wrote:
Carlos E. R. wrote:
See that process using 0,010t of resident memory? I've never seen the word "teras" of ram used for a single process... :-o
I have occasionally seen some crazy numbers in top, no relation to reality.
I have another tab that Firefox about:performance says it is using 2.33GB of ram. A train information page:
<https://www.thetrainline.com/en/train-times/seville-to-berlin>
I have just loaded it, 28Mb.
And now, approx 90mins later, it has grown to 39Mb.
Maybe some ads playing nasty? I have a windows machine here that sometimes crashes the ghostery plugin in the browser. Have to kill the chrome process and reload the plugin.
It is certainly possible. I use uBlock with almost default settings. I do not like to block adds, only "bad" adds. Intrusive adds. -- Cheers / Saludos, Carlos E. R. (from 15.2 x86_64 at Telcontar)
On 24/03/2021 12.53, Per Jessen wrote:
Carlos E. R. wrote:
See that process using 0,010t of resident memory? I've never seen the word "teras" of ram used for a single process... :-o
I have occasionally seen some crazy numbers in top, no relation to reality.
I have another tab that Firefox about:performance says it is using 2.33GB of ram. A train information page:
<https://www.thetrainline.com/en/train-times/seville-to-berlin>
I have just loaded it, 28Mb.
Well, mine had been loaded for days, perhaps weeks. -- Cheers / Saludos, Carlos E. R. (from 15.2 x86_64 at Telcontar)
On 3/23/21 8:52 PM, Carlos E. R. wrote:
what a lot of crap, how can they design those pages that eat so many resources.
You are singing to the choir. All the webbies out there (kids with crayons) are not programmers and have no concept of stack, heap or .bss, .data, .rodata, or .text or transmission protocols, packets, etc... They have GUI's where they drag one widget to a position on a page, click "gradient fill" because the think it looks cool, and a few additional <dim> statements for forms with drop-shadows, add 3 or 4 third-party sites to load javascript animations for the widgets on the forms, and then add google-analytics, qualtrics, etc.. and then type the all important content: "Hello World!" and save to the web-server. Then they wonder why in the hell it takes 2Gig of virtual memory to load their "Hello World" page -- but they are not programmers, so they assume it is normal and move on to adding more content..... The browser programmers then have to load all this crap and make it appear to be responsive to their page-load algorithms happily loop realloc()'ing memory until all the crap has loaded. A good feature would be a virtual memory limit per-page or per-site that may slow the load a bit but would limit pages to 10M. -- David C. Rankin, J.D.,P.E.
On 3/25/21 12:12 AM, David C. Rankin wrote:
On 3/23/21 8:52 PM, Carlos E. R. wrote:
what a lot of crap, how can they design those pages that eat so many resources.
The browser programmers then have to load all this crap and make it appear to be responsive to their page-load algorithms happily loop realloc()'ing memory until all the crap has loaded. A good feature would be a virtual memory limit per-page or per-site that may slow the load a bit but would limit pages to 10M.
Thankfully, my worst offender is ~2.6M: 3231 david 20 0 2626116 144420 100288 S 0.333 1.781 0:58.64 Web Content Thank God for uBlock Origins and NoScript..... If a site doesn't work without all the cruft, it's a site I don't visit. -- David C. Rankin, J.D.,P.E.
On 25/03/2021 06.17, David C. Rankin wrote:
On 3/25/21 12:12 AM, David C. Rankin wrote:
On 3/23/21 8:52 PM, Carlos E. R. wrote:
what a lot of crap, how can they design those pages that eat so many resources.
The browser programmers then have to load all this crap and make it appear to be responsive to their page-load algorithms happily loop realloc()'ing memory until all the crap has loaded. A good feature would be a virtual memory limit per-page or per-site that may slow the load a bit but would limit pages to 10M.
Thankfully, my worst offender is ~2.6M:
3231 david 20 0 2626116 144420 100288 S 0.333 1.781 0:58.64 Web Content
Thank God for uBlock Origins and NoScript..... If a site doesn't work without all the cruft, it's a site I don't visit.
I can't "not visit" sites. I need to find information wherever it is. Years ago (year 2000 or so) I had to replace my computer with a bigger one because I needed to load certain pages that would take minutes to load. And worse, I had to interact with them (job search pages) and wait for the result of my actions (even half an hour). I could not increase the ram, I had to buy a new bigger and faster computer to cope. And this has happened more times. This computer has 32 gigs and I intend to buy another 32 - not because I need them now, but because I will need them in years to come and then perhaps the modules will not be available. (If you want to see a tool that eats memory by truck loads, try to recover an XFS filesystem with lots of metadata. You may need terabytes. Swap can be useful just that day) -- Cheers / Saludos, Carlos E. R. (from 15.2 x86_64 at Telcontar)
On 3/25/21 12:17 AM, David C. Rankin wrote:
On 3/25/21 12:12 AM, David C. Rankin wrote:
On 3/23/21 8:52 PM, Carlos E. R. wrote:
what a lot of crap, how can they design those pages that eat so many resources.
The browser programmers then have to load all this crap and make it appear to be responsive to their page-load algorithms happily loop realloc()'ing memory until all the crap has loaded. A good feature would be a virtual memory limit per-page or per-site that may slow the load a bit but would limit pages to 10M.
Thankfully, my worst offender is ~2.6M:
3231 david 20 0 2626116 144420 100288 S 0.333 1.781 0:58.64 Web Content
Thank God for uBlock Origins and NoScript..... If a site doesn't work without all the cruft, it's a site I don't visit.
Okay, Firefox got me to with the fictitious virtual memory ballooning: 3358 david 20 0 26.648g 173920 105760 S 0.332 2.144 1:41.93 WebExtensions No over-reservation here... only 26.648g on an 8g laptop... -- David C. Rankin, J.D.,P.E.
On 2021-03-25 00:12:55 David C. Rankin wrote:
|On 3/23/21 8:52 PM, Carlos E. R. wrote: |> what a lot of crap, how can they design those pages that eat so many |> resources. | |You are singing to the choir. | |All the webbies out there (kids with crayons) are not programmers and have | no concept of stack, heap or .bss, .data, .rodata, or .text or | transmission protocols, packets, etc... They have GUI's where they drag | one widget to a position on a page, click "gradient fill" because the | think it looks cool, and a few additional <dim> statements for forms with | drop-shadows, add 3 or 4 third-party sites to load javascript animations | for the widgets on the forms, and then add google-analytics, qualtrics, | etc.. and then type the all important content: | | "Hello World!" | |and save to the web-server. | |Then they wonder why in the hell it takes 2Gig of virtual memory to load | their "Hello World" page -- but they are not programmers, so they assume | it is normal and move on to adding more content..... | |The browser programmers then have to load all this crap and make it appear | to be responsive to their page-load algorithms happily loop realloc()'ing | memory until all the crap has loaded. A good feature would be a virtual | memory limit per-page or per-site that may slow the load a bit but would | limit pages to 10M.
Some of those webbies have been given mandates to create the many, many enterprise business web pages that look so very flashy but are difficult-to-impossible to navigate, and that suck up our bandwidth resources to provide little beyond frustration for their customers. Leslie -- openSUSE Leap 15.2 x86_64
On 3/25/21 7:19 AM, J Leslie Turriff wrote:
Some of those webbies have been given mandates to create the many, many enterprise business web pages that look so very flashy but are difficult-to-impossible to navigate, and that suck up our bandwidth resources to provide little beyond frustration for their customers.
I found just that same description under the "urban" dictionary's definition of "progress in the 21st century"... -- David C. Rankin, J.D.,P.E.
participants (8)
-
Carlos E. R. -
Dave Howorth -
David C. Rankin -
David T-G -
J Leslie Turriff -
L A Walsh -
Leen de Braal -
Per Jessen