[chromium-dev] Re: [linux] green flashes in page content
On Tue, May 19, 2009 at 5:12 PM, James Hawkins jhawk...@chromium.org wrote: On Tue, May 19, 2009 at 5:07 PM, Evan Stade est...@chromium.org wrote: Are we sure we need this functionality? I understand the intention of seeing when we're not drawing fast enough, but I'm seeing the green background all the time (on a fast machine), which leads me to believe that the debugging aid is a red herring. Yea, you should see it all the time. We'll never be able to fix that. But how much of it you see is a good measure of jank. If enough people hate it, we can remove it (and only enable when we care to look at jank). We should leave page load/rendering time to measurements from the tests. This should be something that a developer working on jankiness could enable himself. It's rather distracting to the developers who don't need to look into jankiness, but do need to run debug builds regularly. I personally like seeing the green background, since I do like seeing how much jank there is. It doesn't bother me at all. That said, I can see how it'd annoy some people. If most people dislike it, then I think we should just make it the same as the release behavior (use a white background), but I think it'd be nice to have a command line flag or a preference or something so I can run with a green or whatever background to aid in debugging jankiness. -- James Hawkins --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: [linux] green flashes in page content
On Tue, May 19, 2009 at 11:20 PM, William Chan (陈智昌) willc...@chromium.org wrote: On Tue, May 19, 2009 at 5:12 PM, James Hawkins jhawk...@chromium.org wrote: On Tue, May 19, 2009 at 5:07 PM, Evan Stade est...@chromium.org wrote: Are we sure we need this functionality? I understand the intention of seeing when we're not drawing fast enough, but I'm seeing the green background all the time (on a fast machine), which leads me to believe that the debugging aid is a red herring. Yea, you should see it all the time. We'll never be able to fix that. But how much of it you see is a good measure of jank. If enough people hate it, we can remove it (and only enable when we care to look at jank). We should leave page load/rendering time to measurements from the tests. This should be something that a developer working on jankiness could enable himself. It's rather distracting to the developers who don't need to look into jankiness, but do need to run debug builds regularly. I personally like seeing the green background, since I do like seeing how much jank there is. It doesn't bother me at all. That said, I can see how it'd annoy some people. If most people dislike it, then I think we should just make it the same as the release behavior (use a white background), but I think it'd be nice to have a command line flag or a preference or something so I can run with a green or whatever background to aid in debugging jankiness. Sure, I think a flag or pref would be ideal. --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: Drawing code on Mac
Thanks Amanda (and Stephen :-)... I finally got to look at this code in the train this morning (after a very relaxing extended long weekend :-)... And I have an idea of how we could do this more easily and hopefully safe enough... I don't think I can do the save/restore trick in PlatformDeviceMac::LoadClippingRegionToCGContext(), because this a static method with no context... But I noticed that it is only called at one place, within BitmapPlatformDeviceMac::BitmapPlatformDeviceMacData::LoadConfig() which is also the only place where LoadTransformToCGContext() is called, so we could simply rely on restore/save there... But is it safe enough? The other alternative, would be to create an image mask in PlatformDeviceMac::LoadClippingRegionToCGContext(), and use CGContextClipToMask() instead of CGContextClipToRect()... But note that ClipToMask was only introduced in 10.4, I know that our development environment only supports 10.5 and up, but do we expect Chromium to run on versions earlier than 10.4? Thanks again! BYE MAD On Sat, May 16, 2009 at 4:55 PM, Amanda Walker ama...@chromium.org wrote: The code to look at is in skia/ext/platform_device_mac.cc:123 and following (PlatformDeviceMac::LoadClippingRegionToCGContext). It calls CGContextClipToRect, but does not reset the clip path first, so we we end up with the behavior Stephen described. The best solution would be to do a save when the device is created and a restore / save pair before loading a new clip path. We'll have to reload the rest of the Skia drawing state back into the CGContext after the restore, of course, but that shouldn't be too bad. We might also be able to define a mask that we draw through (and that we can reset at will), rather than manipulating the CGContext's clip region at all. It is possible to reset a CGContext's clip region (which matches the Skia drawing model a bit better) using ClipCGContextToRegion, but that was deprecated in 10.4 and won't be around for 64-bit applications, so I'd rather not use it unless the other approaches turns out to be impractical for some reason. --Amanda On Sat, May 16, 2009 at 3:44 PM, Marc-Andre Decoste m...@google.com wrote: Salut Stephen, thanks for the hint... Note that I have tried calling save/restore on the canvas around each sub-rect drawing... With no luck... I also tried to make the call to canvas-getTopPlatformDevice().accessBitmap(false); after every sub-rect... same thing... I guess I'll wait for Amanda to take a look at it tomorrow or Monday... I quickly looked at the skia/ext/platform_canvas_mac.cc file content but didn't see anything related to clip rects in there... Thanks! BYE MAD On Sat, May 16, 2009 at 2:13 PM, Stephen White senorbla...@chromium.org wrote: On Sat, May 16, 2009 at 10:16 AM, Marc-Andre Decoste m...@google.com wrote: Salut, I found the cause of the problem, but didn't find a cure yet.. The SkCanvas::ClipRect() call with the replace Op doesn't seem to work more than once... The first rect gets drawn on the cancas properly, but all subsequent rects drawn on the same canvas with different clip rects seem to be ignored... If I don't make the call to clip the rect, then all the rects are drawn, but some of the earlier rects can be over-painted with white when other rects are drawn at the same height within the canvas... I tried to clear the clip rect after drawing each rectagle (using the kDifference_Op), also tried unioning the clip rect before doing an intersect op, still with no success... I'll keep digging to understand why the subsequent clip rects don't work on the Mac when they work fine on Windows and Linux... But I would appreciate some hints if you have some... :-) I haven't looked at the code, but this sounds a lot like a peculiarity of CoreGraphics I ran into a while back: setting the clip path with CGContextClipToRect() does the intersection of that rect with the existing clip path, rather than replacing it. Since the clip path can't be made larger, the only way to really restore it is to save the graphics state before setting the clip path, and restore the graphics state after drawing. So if we're not doing that, it's possible that the CoreGraphics path may not have the same semantics as the Skia path in this case. (It's my understanding that we use CoreGraphics on the Mac rather than Skia, but since you mentioned SkCanvas, I could be wrong.) Stephen On Wed, May 13, 2009 at 1:25 PM, Marc-Andre Decoste m...@google.com wrote: As I mentioned in the latest message, this code is OK after all :-)... The problem seems to be at the other end, when I split the drawing in sub-regions of the TranportDIB, on the rendering side... I wasn't expecting this, since this other code is not using any platform specific stuff directly... You can find it in here:
[chromium-dev] VS2008 dependencies..... something is wrong....
how does one see which project depends on which ? i ask this because today i had a problem so i clobbered my debug folder, and out of curiosity went to chrome.exe and did build this project only and it did not even bother to compile all the projects chrome.exe is based on! also, even a full re-build did not solve my problems and i still had 2 missing libs 'views.lib' and 'utility.lib' which i had to go to manually and build them how are dependencies expressed inside VS2008 regarding chrome ? --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: Drawing code on Mac
On Wed, May 20, 2009 at 9:22 AM, Marc-Andre Decoste m...@google.com wrote: I don't think I can do the save/restore trick in PlatformDeviceMac::LoadClippingRegionToCGContext(), because this a static method with no context... But I noticed that it is only called at one place, within BitmapPlatformDeviceMac::BitmapPlatformDeviceMacData::LoadConfig() which is also the only place where LoadTransformToCGContext() is called, so we could simply rely on restore/save there... But is it safe enough? Hmm, it should be. The purpose of BitmapPlatformDeviceMacData::LoadConfig is to sync up the native graphics context with the PlatformDevice settings. Since the model that Skia and Windows use is set rather than a stack, a restore/save to go back to the defaults before loading up the current settings seems completely reasonable. The other alternative, would be to create an image mask in PlatformDeviceMac::LoadClippingRegionToCGContext(), and use CGContextClipToMask() instead of CGContextClipToRect()... But note that ClipToMask was only introduced in 10.4, I know that our development environment only supports 10.5 and up, but do we expect Chromium to run on versions earlier than 10.4? Earlier than 10.4 is not a concern at all, so anything introduced in 10.4 is fair game. I would expect using a mask to use more memory than a list of rectangles, so if the restore/save works, it's probably worth trying that first. Of course, clipping to a mask may be faster, so ... shrug. Use your best judgement :-). --Amanda --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: Drawing code on Mac
The quick and easy restore/save trick in LoadConfig() seems to work fine (at least in all the tests I did)... I'll update the patch with these change and publish it for review... Let me know ASAP if you think this is a bad idea so that I can try the other (more complex, though probably safer) solution using the mask.. Thanks! BYE MAD On Wed, May 20, 2009 at 9:22 AM, Marc-Andre Decoste m...@google.com wrote: Thanks Amanda (and Stephen :-)... I finally got to look at this code in the train this morning (after a very relaxing extended long weekend :-)... And I have an idea of how we could do this more easily and hopefully safe enough... I don't think I can do the save/restore trick in PlatformDeviceMac::LoadClippingRegionToCGContext(), because this a static method with no context... But I noticed that it is only called at one place, within BitmapPlatformDeviceMac::BitmapPlatformDeviceMacData::LoadConfig() which is also the only place where LoadTransformToCGContext() is called, so we could simply rely on restore/save there... But is it safe enough? The other alternative, would be to create an image mask in PlatformDeviceMac::LoadClippingRegionToCGContext(), and use CGContextClipToMask() instead of CGContextClipToRect()... But note that ClipToMask was only introduced in 10.4, I know that our development environment only supports 10.5 and up, but do we expect Chromium to run on versions earlier than 10.4? Thanks again! BYE MAD On Sat, May 16, 2009 at 4:55 PM, Amanda Walker ama...@chromium.orgwrote: The code to look at is in skia/ext/platform_device_mac.cc:123 and following (PlatformDeviceMac::LoadClippingRegionToCGContext). It calls CGContextClipToRect, but does not reset the clip path first, so we we end up with the behavior Stephen described. The best solution would be to do a save when the device is created and a restore / save pair before loading a new clip path. We'll have to reload the rest of the Skia drawing state back into the CGContext after the restore, of course, but that shouldn't be too bad. We might also be able to define a mask that we draw through (and that we can reset at will), rather than manipulating the CGContext's clip region at all. It is possible to reset a CGContext's clip region (which matches the Skia drawing model a bit better) using ClipCGContextToRegion, but that was deprecated in 10.4 and won't be around for 64-bit applications, so I'd rather not use it unless the other approaches turns out to be impractical for some reason. --Amanda On Sat, May 16, 2009 at 3:44 PM, Marc-Andre Decoste m...@google.com wrote: Salut Stephen, thanks for the hint... Note that I have tried calling save/restore on the canvas around each sub-rect drawing... With no luck... I also tried to make the call to canvas-getTopPlatformDevice().accessBitmap(false); after every sub-rect... same thing... I guess I'll wait for Amanda to take a look at it tomorrow or Monday... I quickly looked at the skia/ext/platform_canvas_mac.cc file content but didn't see anything related to clip rects in there... Thanks! BYE MAD On Sat, May 16, 2009 at 2:13 PM, Stephen White senorbla...@chromium.org wrote: On Sat, May 16, 2009 at 10:16 AM, Marc-Andre Decoste m...@google.com wrote: Salut, I found the cause of the problem, but didn't find a cure yet.. The SkCanvas::ClipRect() call with the replace Op doesn't seem to work more than once... The first rect gets drawn on the cancas properly, but all subsequent rects drawn on the same canvas with different clip rects seem to be ignored... If I don't make the call to clip the rect, then all the rects are drawn, but some of the earlier rects can be over-painted with white when other rects are drawn at the same height within the canvas... I tried to clear the clip rect after drawing each rectagle (using the kDifference_Op), also tried unioning the clip rect before doing an intersect op, still with no success... I'll keep digging to understand why the subsequent clip rects don't work on the Mac when they work fine on Windows and Linux... But I would appreciate some hints if you have some... :-) I haven't looked at the code, but this sounds a lot like a peculiarity of CoreGraphics I ran into a while back: setting the clip path with CGContextClipToRect() does the intersection of that rect with the existing clip path, rather than replacing it. Since the clip path can't be made larger, the only way to really restore it is to save the graphics state before setting the clip path, and restore the graphics state after drawing. So if we're not doing that, it's possible that the CoreGraphics path may not have the same semantics as the Skia path in this case. (It's my understanding that we use CoreGraphics on the Mac rather than Skia, but since you mentioned SkCanvas, I could be wrong.) Stephen On
[chromium-dev] Re: VS2008 dependencies..... something is wrong....
VS 2008 pro SP 1, on Vista (hebrew) 32SP1 my other projects works just fine, i always thought that the dependencies don't show because you use some kind of strange build system --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: VS2008 dependencies..... something is wrong....
also, mark please have a look at this consistent unit_tests bug i got no response, and this makes me have to log out and log in time and again http://code.google.com/p/chromium/issues/detail?id=12093can=4sort=-idcolspec=ID%20Stars%20Pri%20Area%20Type%20Status%20Summary%20Modified%20Owner%20Mstone --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: VS2008 dependencies..... something is wrong....
mark, i plan to contrib code to chrome so i wanted to run all the unit tests (also of the UI) in the last 2 days all my builds return link errors .since i switched to safesync_url: http://build.chromium.org/buildbot/continuous/ LATEST/REVISION what is worse is that the IDE seems to create chrome_dll and the rest without compiling all the files after a sync... this is why i asked how does VS know which dependencies a project relies on normally, it is easy to see via the project-dependencies but in chrome case those menus are greyed out --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: VS2008 dependencies..... something is wrong....
On Wed, May 20, 2009 at 11:38 AM, nakro yoav.zilberb...@gmail.com wrote: normally, it is easy to see via the project-dependencies but in chrome case those menus are greyed out Something is wrong in your VS setup. What version (2005, 2008) and flavor (Express, Standard, Pro, Trial, etc) of Visual Studio are you using? M-A --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: VS2008 dependencies..... something is wrong....
S:\chrome\n1\src\chromepython tools\build\win\sln_deps.py chrome.sln chrome_exe chrome_exe base breakpad_handler chrome_dll crash_service en-US sandbox theme_dll util chrome.exe shouldn't depend on views BTW. M-A On Wed, May 20, 2009 at 9:31 AM, nakro yoav.zilberb...@gmail.com wrote: how does one see which project depends on which ? i ask this because today i had a problem so i clobbered my debug folder, and out of curiosity went to chrome.exe and did build this project only and it did not even bother to compile all the projects chrome.exe is based on! also, even a full re-build did not solve my problems and i still had 2 missing libs 'views.lib' and 'utility.lib' which i had to go to manually and build them how are dependencies expressed inside VS2008 regarding chrome ? --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: [linux] green flashes in page content
I think it would be great if I could turn on a pref in release builds that would show me jank. I'd also like to see all of the UI elements turn crazy colors if we fail to service the event queue in less than 200ms. Linus 2009/5/19 James Hawkins jhawk...@chromium.org On Tue, May 19, 2009 at 11:20 PM, William Chan (陈智昌) willc...@chromium.org wrote: On Tue, May 19, 2009 at 5:12 PM, James Hawkins jhawk...@chromium.org wrote: On Tue, May 19, 2009 at 5:07 PM, Evan Stade est...@chromium.org wrote: Are we sure we need this functionality? I understand the intention of seeing when we're not drawing fast enough, but I'm seeing the green background all the time (on a fast machine), which leads me to believe that the debugging aid is a red herring. Yea, you should see it all the time. We'll never be able to fix that. But how much of it you see is a good measure of jank. If enough people hate it, we can remove it (and only enable when we care to look at jank). We should leave page load/rendering time to measurements from the tests. This should be something that a developer working on jankiness could enable himself. It's rather distracting to the developers who don't need to look into jankiness, but do need to run debug builds regularly. I personally like seeing the green background, since I do like seeing how much jank there is. It doesn't bother me at all. That said, I can see how it'd annoy some people. If most people dislike it, then I think we should just make it the same as the release behavior (use a white background), but I think it'd be nice to have a command line flag or a preference or something so I can run with a green or whatever background to aid in debugging jankiness. Sure, I think a flag or pref would be ideal. --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: VS2008 dependencies..... something is wrong....
bradley, if i open the src/base/base.sln as the solution i see dependencies, for base_unittest (or something like this) but in all my projects (i am a developer) i always see dependencies, chromium is hardly my first project. and an i not just petty about this, the build goes wrong after i do a sync as some modules do not get compiled so at best i get link error and at worse (happened a few times) i get an exe who crashes because of different struct sizes am i the only who who has the dependencies greyed out for chrome.sln ? --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: VS2008 dependencies..... something is wrong....
Do you see dependencies in any of the slns (like src/base/base.sln for instance?) -BradN On Wed, May 20, 2009 at 9:30 AM, nakro yoav.zilberb...@gmail.com wrote: VS 2008 pro SP 1, on Vista (hebrew) 32SP1 my other projects works just fine, i always thought that the dependencies don't show because you use some kind of strange build system --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: [linux] green flashes in page content
The mac build does something like this: the background color of all the windows is actually magenta, which makes it very obvious when something's not getting painted or the process isn't responding to events. --Amanda 2009/5/20 Linus Upson li...@google.com: I think it would be great if I could turn on a pref in release builds that would show me jank. I'd also like to see all of the UI elements turn crazy colors if we fail to service the event queue in less than 200ms. Linus 2009/5/19 James Hawkins jhawk...@chromium.org On Tue, May 19, 2009 at 11:20 PM, William Chan (陈智昌) willc...@chromium.org wrote: On Tue, May 19, 2009 at 5:12 PM, James Hawkins jhawk...@chromium.org wrote: On Tue, May 19, 2009 at 5:07 PM, Evan Stade est...@chromium.org wrote: Are we sure we need this functionality? I understand the intention of seeing when we're not drawing fast enough, but I'm seeing the green background all the time (on a fast machine), which leads me to believe that the debugging aid is a red herring. Yea, you should see it all the time. We'll never be able to fix that. But how much of it you see is a good measure of jank. If enough people hate it, we can remove it (and only enable when we care to look at jank). We should leave page load/rendering time to measurements from the tests. This should be something that a developer working on jankiness could enable himself. It's rather distracting to the developers who don't need to look into jankiness, but do need to run debug builds regularly. I personally like seeing the green background, since I do like seeing how much jank there is. It doesn't bother me at all. That said, I can see how it'd annoy some people. If most people dislike it, then I think we should just make it the same as the release behavior (use a white background), but I think it'd be nice to have a command line flag or a preference or something so I can run with a green or whatever background to aid in debugging jankiness. Sure, I think a flag or pref would be ideal. --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: VS2008 dependencies..... something is wrong....
one more thing which could be different on my machine i did set GYP_MSVS_VERSION=2008 but i doubt this could be it still, (and i have no VS2005 or 2010 installed, so it should have been this anyways) --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: cygwin setup_mount does not run automatically
Ah my change to fix that got blown away with other chrome.sln changes.I've re-landed it. Let me know if you see this anymore. -BradN On Wed, May 20, 2009 at 1:29 PM, Miranda Callahan miran...@google.comwrote: here's the full data, if it helps: Microsoft Visual Studio 2005 Version 8.0.50727.867 (vsvista.050727-8600) Microsoft .NET Framework Version 2.0.50727 SP1 Installed Edition: Professional Microsoft Visual Basic 2005 77626-009-007-41731 Microsoft Visual Basic 2005 Microsoft Visual C# 2005 77626-009-007-41731 Microsoft Visual C# 2005 Microsoft Visual C++ 2005 77626-009-007-41731 Microsoft Visual C++ 2005 Microsoft Visual J# 2005 77626-009-007-41731 Microsoft Visual J# 2005 Microsoft Visual Web Developer 2005 77626-009-007-41731 Microsoft Visual Web Developer 2005 Microsoft Web Application Projects 2005 77626-009-007-41731 Microsoft Web Application Projects 2005 Version 8.0.50727.867 Crystal ReportsAAC60-G0CSA4B-V7000AY Crystal Reports for Visual Studio 2005 Hotfix for Microsoft Visual Studio 2005 Professional Edition - ENU (KB935225) This Hotfix is for Microsoft Visual Studio 2005 Professional Edition - ENU. If you later install a more recent service pack, this Hotfix will be uninstalled automatically. For more information, visit http://support.microsoft.com/kb/935225 Microsoft Visual Studio 2005 Professional Edition - ENU Service Pack 1 (KB926601) This service pack is for Microsoft Visual Studio 2005 Professional Edition - ENU. If you later install a more recent service pack, this service pack will be uninstalled automatically. For more information, visit http://support.microsoft.com/kb/926601 Update for Microsoft Visual Studio 2005 Professional Edition - ENU (KB932232) This Update is for Microsoft Visual Studio 2005 Professional Edition - ENU. If you later install a more recent service pack, this Update will be uninstalled automatically. For more information, visit http://support.microsoft.com/kb/932232 On Wed, May 20, 2009 at 1:28 PM, Miranda Callahan miran...@google.comwrote: Hi, Brad -- nope, VS 2005, sp1... - Miranda On Wed, May 20, 2009 at 1:25 PM, Bradley Nelson bradnel...@google.comwrote: Hi Miranda, Are you using vs2008 by any chance? -BradN On Wed, May 20, 2009 at 1:07 PM, Miranda Callahan miran...@google.comwrote: Hi Brad -- I found this old link in chromium-dev that explained what to do -- I had to manually set the cygwin registry vars. For some reason, it wasn't being automatically set by the build. http://groups.google.com/group/chromium-dev/browse_thread/thread/20833bff27dddffe/28b94d11c56f449e?lnk=gstq=cygwin+setup_mount#28b94d11c56f449e http://groups.google.com/group/chromium-dev/browse_thread/thread/20833bff27dddffe/28b94d11c56f449e?lnk=gstq=cygwin+setup_mount#28b94d11c56f449e- Miranda On Wed, May 20, 2009 at 11:54 AM, Miranda Callahan miran...@google.com wrote: The whole sln -- just got a new machine, got the tarball, gclient sync --force -- and trying to build chrome.sln. I was wondering if it could be interference between my cygwin, and the cygwin commands in the build? On Wed, May 20, 2009 at 11:45 AM, Bradley Nelson bradnel...@google.com wrote: Are you building a particular subproject, or the whole sln? -BradN On Wed, May 20, 2009 at 11:44 AM, miran...@chromium.org miran...@chromium.org wrote: Hi Brad, I'm getting the same error when trying a clean build on Vista 64 -- any clues how I can fix it? Thanks! - Miranda On May 16, 5:05 pm, Bradley Nelson bradnel...@google.com wrote: Hi Ben, I've just checked in a change that should cause that to get run at the right time. Please let me know if you see this issue again. -BradN On Fri, May 15, 2009 at 2:22 AM, Ben Harper rogo...@gmail.com wrote: Hi, I don't know whether cygwin/setup_mount is supposed to run automatically at some stage of the checkout/build process, but for me it did not, and it took me a while to figure out that it was necessary. Initial symptom was this error: Can't locate strict.pm in @INC when trying to run dftables. Just thought I'd mention it here. --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: Bookmark system not reliable enough for user preferences
On Tue, May 19, 2009 at 11:35 AM, jack js2...@gmail.com wrote: Sorry for this late response. I was piled up in the past days. I did some additional trials. It's quite random and it is hard to get unique results. But let's handle it one by one. I am attaching a sample toolstrip.html file. For a clean test make sure your bookmark list is empty. When the page is loaded it will check the existence of a folder and create it (together w/ an item) if needed. In my test, the item will never be created in the folder. It will be either created in the bookmark bar, or under the root folder, or not at all. You can also click on the icon in the toolstrip to manually check the folder. In this case, I got correct result in one machine, but random (as aforementioned) in another machine (note this another machine is using a different language version). Also, a folder is never found by chrome.bookmarks.search(). It is easy for you to test it so I didn't include it in the html file. Hi Jack, I can repro this. The problem is a race at startup. If you delay calling the APIs awhile, it would work correctly. I've created a bug for this, if you're interested: http://code.google.com/p/chromium/issues/detail?id=12353 As for searching for bookmark folders, I'm not sure we want to add it. If it wasn't for this weird use case of using the bookmark system as storage, I don't know if anybody would ever need that. Hopefully real storage appears soon, and we can use that instead. - a --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: Bookmark system not reliable enough for user preferences
On Wed, May 20, 2009 at 2:54 PM, Aaron Boodman a...@chromium.org wrote: On Tue, May 19, 2009 at 11:35 AM, jack js2...@gmail.com wrote: Sorry for this late response. I was piled up in the past days. I did some additional trials. It's quite random and it is hard to get unique results. But let's handle it one by one. I am attaching a sample toolstrip.html file. For a clean test make sure your bookmark list is empty. When the page is loaded it will check the existence of a folder and create it (together w/ an item) if needed. In my test, the item will never be created in the folder. It will be either created in the bookmark bar, or under the root folder, or not at all. You can also click on the icon in the toolstrip to manually check the folder. In this case, I got correct result in one machine, but random (as aforementioned) in another machine (note this another machine is using a different language version). Also, a folder is never found by chrome.bookmarks.search(). It is easy for you to test it so I didn't include it in the html file. Hi Jack, I can repro this. The problem is a race at startup. If you delay calling the APIs awhile, it would work correctly. I've created a bug for this, if you're interested: http://code.google.com/p/chromium/issues/detail?id=12353 As for searching for bookmark folders, I'm not sure we want to add it. If it wasn't for this weird use case of using the bookmark system as storage, I don't know if anybody would ever need that. Hopefully real storage appears soon, and we can use that instead. I'm currently doing some refactoring work on the WebKit LocalStorage implementation so that it can fit into Chromium's multi-processs model. It's hard to say exactly how long it'll take, but I'm definitely working on it. :-) I think Dumi's pretty far along on Databases as well. --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: Bookmark system not reliable enough for user preferences
On Wed, May 20, 2009 at 3:03 PM, Jeremy Orlow jor...@chromium.org wrote: Hopefully real storage appears soon, and we can use that instead. I'm currently doing some refactoring work on the WebKit LocalStorage implementation so that it can fit into Chromium's multi-processs model. It's hard to say exactly how long it'll take, but I'm definitely working on it. :-) I think Dumi's pretty far along on Databases as well. Thanks Jeremy, I know you and Dumi are working hard on it. Not trying to pester. - a --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: Bookmark system not reliable enough for user preferences
Thanks for all of your input. Your finding confirmed my suspicion about racing, which might also be the reason for several other weird random behaviors I encountered. Looking forward to seeing any progress in LocalStorage, but take your time. For me it is interesting to hacking into current release and understand the way it cooks. -Jack On May 20, 2:54 pm, Aaron Boodman a...@chromium.org wrote: Hi Jack, I can repro this. The problem is a race at startup. If you delay calling the APIs awhile, it would work correctly. I've created a bug for this, if you're interested:http://code.google.com/p/chromium/issues/detail?id=12353 As for searching for bookmark folders, I'm not sure we want to add it. If it wasn't for this weird use case of using the bookmark system as storage, I don't know if anybody would ever need that. Hopefully real storage appears soon, and we can use that instead. - a --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Nomination for Chromium Account
Hi, I'm Mohamed, and been working on chromium for some time now (windows and now Linux), and fixing chromium bugs to learn the C++ language and actually being a part of something important. Knowing that anything I do, would affect many users around the globe, which is an awesome feeling that I can't do in my day job. I graduated 6 month before I started Chromium, and I work full time as a Software Engineer doing Java Enterprise development (ui, security, core). After work/gym, I have little time so I really enjoy sitting down and working with Chromium, its very enjoyable! I wish it is full time, but my boss needs me in the Java world! I am very active in Google Help Chrome Forums (where I am a Top Contributor), mailing lists (discuss, and dev), and of course IRC (IRC fan since 1999 when I was 12), and a fan in the real world, showing it off anywhere I go. I came in fresh without knowing anything about python, c++, lieux devel, and chromium/webkit code base. I quite learned a lot which I am grateful, thanks to pkasting being very strict on everything, maruel being very helpful on Google chat/irc, hunar helping with problems late at night (1AM+), and all the IRC folks who make this enjoyable (evanm should sleep more :p he is always around)! I am learning the correct way developing applications, the powerful software processes, and best of all... gdb is so darn cool (learning it now). I got into Linux development 2 days ago and already submitted Cl's (first time doing Gtk and Linux dev and hope to continue cause so nice). I uninstalled windows just to install Linux and code chromium :) (As well forcing myself to stop playing games, of course.). I know that I am not that experienced in this type of programming and environment (but I believe I am now), and I have solved some tough problems, but I would like to have a chromium account. That is the only way I could use the try-servers! I don't know if I am qualified yet, but I believe I am :). So far I haven't submitted that much since January (I was on break for a month and half, which explains the lack of bug fixes). But out of my 30+ CL's only: * 22 committed * 9 still opened (waiting for commit) * ~5 from issue tracker cancelled due to not needed or closed since its old (no one committed) If any other information is needed, please let me know. Thanks! Here are my 22 committed trackers. -[ 1 ]-- Revision: r16360 Comitter: aba...@chromium.org Date: 2009-05-19 Files Changed: 1 Lines Changed: 11 Note: Fix Linux Bustage from new CL where it is missing string import CR: http://codereview.chromium.org/113573 -[ 2 ]-- Revision: r15220 Comitter: s...@chromium.org Date: 2009-05-04 Files Changed: 2 Lines Changed: 21 Note: Mouse hover out of context menu should un-select menu item. CR: http://codereview.chromium.org/100022 -[ 3 ]-- Revision: r14731 Comitter: mar...@chromium.org Date: 2009-04-28 Files Changed: 3 Lines Changed: 13 Note: Should handle the events when no printer is detected on local machine CR: http://codereview.chromium.org/100023 -[ 4 ]-- Revision: r14688 Comitter: e...@chromium.org Date: 2009-04-27 Files Changed: 10 Lines Changed: 9 Note: Convert Windows Sleep to PlatformThread::Sleep CR: http://codereview.chromium.org/63113 -[ 5 ]-- Revision: r14653 Comitter: t...@google.com Date: 2009-04-27 Files Changed: 1 Lines Changed: 12 Note: Update the build installation script to make it work on ubuntu 9.04 CR: http://codereview.chromium.org/92155 -[ 6 ]-- Revision: r14288 Comitter: hb...@chromium.org Date: 2009-04-23 Files Changed: 1 Lines Changed: 16 Note: Browser Crash when opening urls with NEW_BACKGROUND_TAB disposition CR: http://codereview.chromium.org/93058 -[ 7 ]-- Revision: r13831 Comitter: hb...@chromium.org Date: 2009-04-16 Files Changed: 4 Lines Changed: 15 Note: Introduce printing test case (Step 1) CR: http://codereview.chromium.org/67202 -[ 8 ]-- Revision: r13616 Comitter: t...@google.com Date: 2009-04-13 Files Changed: 1 Lines Changed: 28 Note: Fix XML Reader crash on reading files. CR: http://codereview.chromium.org/67094 -[ 9 ]-- Revision: r13426 Comitter: mar...@chromium.org Date: 2009-04-09 Files Changed: 5 Lines Changed: 52 Note: Implement Printer Preference Persistence (Step 1/3) CR: http://codereview.chromium.org/60118 -[ 10 ]-- Revision: r13177 Comitter: s...@chromium.org Date: 2009-04-06 Files Changed: 1 Lines Changed: 7 Note: The bookmarks title in the bubble should be left justified CR: http://codereview.chromium.org/62067 -[ 11 ]-- Revision: r13045 Comitter: est...@chromium.org Date: 2009-04-02 Files Changed: 4 Lines Changed: 6 Note: Only create one downloads tab per window. Introduce FOCUS_SINGLETON_TAB
[chromium-dev] Re: Tab-modal dialogs on the Mac
Avi, this is cool enough that you should do a blog post about it once landed. jrg On Wed, May 20, 2009 at 3:06 PM, Avi Drissman a...@google.com wrote: OK, this fixes just about everything important. Known issues: - Expose funkiness with 10.5. Since that's fixed in 10.6, that's unlikely to get attention. - Window control funkiness, where they appear disabled but aren't. I don't think that's huge. - No shadow at the top. Mark cares more than I do. Please pound on this hard. I have a DTS incident in (not that I got any useful advice from it) and I'd like to get back to him within, say, a week on the off-chance that we find anything and he's willing to give me a hand. Avi --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Mozilla announces Jetpack extension initiative
https://jetpack.mozillalabs.com/ Similar-sounding goals to Chrome extensions: web-tech-based, no restart needed, backwards-compatible, etc. Not sure how much overlap there is between the two. Any Mozillians who read this (hi beltzner!) feel free to chime in. It would be very cool if these would run in Chrome and Chrome extensions in Firefox. PK --~--~-~--~~~---~--~~ Chromium Developers mailing list: chromium-dev@googlegroups.com View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~--~~~~--~~--~--~---
[chromium-dev] Re: cygwin setup_mount does not run automatically
the latest checkout solving the problem of locating strict.pm, but another problem shows up: 1-- Build started: Project: cygwin, Configuration: Debug Win32 -- 1setup_mount 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1Build log was saved at file://C:\Users\miletus\chrome\chrome\src \chrome\Debug\obj\cygwin\BuildLog.htm 1cygwin - 0 error(s), 0 warning(s) 2-- Build started: Project: pcre, Configuration: Debug Win32 -- 2dftables 2Can't locate File/Basename.pm in @INC (@INC contains: /usr/lib/ perl5/5.10/i686-cygwin /usr/lib/perl5/5.10 /usr/lib/perl5/site_perl/ 5.10/i686-cygwin /usr/lib/perl5/site_perl/5.10 /usr/lib/perl5/ vendor_perl/5.10/i686-cygwin /usr/lib/perl5/vendor_perl/5.10 /usr/lib/ perl5/vendor_perl/5.10 /usr/lib/perl5/site_perl/5.8 /usr/lib/perl5/ vendor_perl/5.8 .) at ../third_party/WebKit/JavaScriptCore/pcre/ dftables line 47. 2BEGIN failed--compilation aborted at ../third_party/WebKit/ JavaScriptCore/pcre/dftables line 47. 2Project : error PRJ0002 : Error result 2 returned from 'C:\Windows \SysWow64\cmd.exe'. 2Build log was saved at file://C:\Users\miletus\chrome\chrome\src \chrome\Debug\obj\pcre\BuildLog.htm 2pcre - 1 error(s), 0 warning(s) any idea ? On May 20, 4:56 pm, Bradley Nelson bradnel...@google.com wrote: Ah my change to fix that got blown away with other chrome.sln changes.I've re-landed it. Let me know if you see this anymore. -BradN On Wed, May 20, 2009 at 1:29 PM, Miranda Callahan miran...@google.comwrote: here's the full data, if it helps: Microsoft Visual Studio 2005 Version 8.0.50727.867 (vsvista.050727-8600) Microsoft .NET Framework Version 2.0.50727 SP1 Installed Edition: Professional Microsoft Visual Basic 2005 77626-009-007-41731 Microsoft Visual Basic 2005 Microsoft Visual C# 2005 77626-009-007-41731 Microsoft Visual C# 2005 Microsoft Visual C++ 2005 77626-009-007-41731 Microsoft Visual C++ 2005 Microsoft Visual J# 2005 77626-009-007-41731 Microsoft Visual J# 2005 Microsoft Visual Web Developer 2005 77626-009-007-41731 Microsoft Visual Web Developer 2005 Microsoft Web Application Projects 2005 77626-009-007-41731 Microsoft Web Application Projects 2005 Version 8.0.50727.867 Crystal Reports AAC60-G0CSA4B-V7000AY Crystal Reports for Visual Studio 2005 Hotfix for Microsoft Visual Studio 2005 Professional Edition - ENU (KB935225) This Hotfix is for Microsoft Visual Studio 2005 Professional Edition - ENU. If you later install a more recent service pack, this Hotfix will be uninstalled automatically. For more information, visithttp://support.microsoft.com/kb/935225 Microsoft Visual Studio 2005 Professional Edition - ENU Service Pack 1 (KB926601) This service pack is for Microsoft Visual Studio 2005 Professional Edition - ENU. If you later install a more recent service pack, this service pack will be uninstalled automatically. For more information, visithttp://support.microsoft.com/kb/926601 Update for Microsoft Visual Studio 2005 Professional Edition - ENU (KB932232) This Update is for Microsoft Visual Studio 2005 Professional Edition - ENU. If you later install a more recent service pack, this Update will be uninstalled automatically. For more information, visithttp://support.microsoft.com/kb/932232 On Wed, May 20, 2009 at 1:28 PM, Miranda Callahan miran...@google.comwrote: Hi, Brad -- nope, VS 2005, sp1... - Miranda On Wed, May 20, 2009 at 1:25 PM, Bradley Nelson bradnel...@google.comwrote: Hi Miranda, Are you using vs2008 by any chance? -BradN On Wed, May 20, 2009 at 1:07 PM, Miranda Callahan miran...@google.comwrote: Hi Brad -- I found this old link in chromium-dev that explained what to do -- I had to manually set the cygwin registry vars. For some reason, it wasn't being automatically set by the build. http://groups.google.com/group/chromium-dev/browse_thread/thread/2083... http://groups.google.com/group/chromium-dev/browse_thread/thread/2083...- Miranda On Wed, May 20, 2009 at 11:54 AM, Miranda Callahan miran...@google.com wrote: The whole sln -- just got a new machine, got the tarball, gclient sync --force -- and trying to build chrome.sln. I was wondering if it could be interference between my cygwin, and the cygwin commands in the build? On Wed, May 20, 2009 at 11:45 AM, Bradley Nelson
[chromium-dev] Re: cygwin setup_mount does not run automatically
I had the same problem with locating strict.pm and now with the latest checkout the problem is solved, but another problem shows up: 1-- Build started: Project: cygwin, Configuration: Debug Win32 -- 1setup_mount 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1The operation completed successfully. 1Build log was saved at file://C:\Users\miletus\chrome\chrome\src \chrome\Debug\obj\cygwin\BuildLog.htm 1cygwin - 0 error(s), 0 warning(s) 2-- Build started: Project: pcre, Configuration: Debug Win32 -- 2dftables 2Can't locate File/Basename.pm in @INC (@INC contains: /usr/lib/ perl5/5.10/i686-cygwin /usr/lib/perl5/5.10 /usr/lib/perl5/site_perl/ 5.10/i686-cygwin /usr/lib/perl5/site_perl/5.10 /usr/lib/perl5/ vendor_perl/5.10/i686-cygwin /usr/lib/perl5/vendor_perl/5.10 /usr/lib/ perl5/vendor_perl/5.10 /usr/lib/perl5/site_perl/5.8 /usr/lib/perl5/ vendor_perl/5.8 .) at ../third_party/WebKit/JavaScriptCore/pcre/ dftables line 47. 2BEGIN failed--compilation aborted at ../third_party/WebKit/ JavaScriptCore/pcre/dftables line 47. 2Project : error PRJ0002 : Error result 2 returned from 'C:\Windows \SysWow64\cmd.exe'. 2Build log was saved at file://C:\Users\miletus\chrome\chrome\src \chrome\Debug\obj\pcre\BuildLog.htm 2pcre - 1 error(s), 0 warning(s) any idea ? On May 20, 4:56 pm, Bradley Nelson bradnel...@google.com wrote: Ah my change to fix that got blown away with other chrome.sln changes.I've re-landed it. Let me know if you see this anymore. -BradN On Wed, May 20, 2009 at 1:29 PM, Miranda Callahan miran...@google.comwrote: here's the full data, if it helps: Microsoft Visual Studio 2005 Version 8.0.50727.867 (vsvista.050727-8600) Microsoft .NET Framework Version 2.0.50727 SP1 Installed Edition: Professional Microsoft Visual Basic 2005 77626-009-007-41731 Microsoft Visual Basic 2005 Microsoft Visual C# 2005 77626-009-007-41731 Microsoft Visual C# 2005 Microsoft Visual C++ 2005 77626-009-007-41731 Microsoft Visual C++ 2005 Microsoft Visual J# 2005 77626-009-007-41731 Microsoft Visual J# 2005 Microsoft Visual Web Developer 2005 77626-009-007-41731 Microsoft Visual Web Developer 2005 Microsoft Web Application Projects 2005 77626-009-007-41731 Microsoft Web Application Projects 2005 Version 8.0.50727.867 Crystal Reports AAC60-G0CSA4B-V7000AY Crystal Reports for Visual Studio 2005 Hotfix for Microsoft Visual Studio 2005 Professional Edition - ENU (KB935225) This Hotfix is for Microsoft Visual Studio 2005 Professional Edition - ENU. If you later install a more recent service pack, this Hotfix will be uninstalled automatically. For more information, visithttp://support.microsoft.com/kb/935225 Microsoft Visual Studio 2005 Professional Edition - ENU Service Pack 1 (KB926601) This service pack is for Microsoft Visual Studio 2005 Professional Edition - ENU. If you later install a more recent service pack, this service pack will be uninstalled automatically. For more information, visithttp://support.microsoft.com/kb/926601 Update for Microsoft Visual Studio 2005 Professional Edition - ENU (KB932232) This Update is for Microsoft Visual Studio 2005 Professional Edition - ENU. If you later install a more recent service pack, this Update will be uninstalled automatically. For more information, visithttp://support.microsoft.com/kb/932232 On Wed, May 20, 2009 at 1:28 PM, Miranda Callahan miran...@google.comwrote: Hi, Brad -- nope, VS 2005, sp1... - Miranda On Wed, May 20, 2009 at 1:25 PM, Bradley Nelson bradnel...@google.comwrote: Hi Miranda, Are you using vs2008 by any chance? -BradN On Wed, May 20, 2009 at 1:07 PM, Miranda Callahan miran...@google.comwrote: Hi Brad -- I found this old link in chromium-dev that explained what to do -- I had to manually set the cygwin registry vars. For some reason, it wasn't being automatically set by the build. http://groups.google.com/group/chromium-dev/browse_thread/thread/2083... http://groups.google.com/group/chromium-dev/browse_thread/thread/2083...- Miranda On Wed, May 20, 2009 at 11:54 AM, Miranda Callahan miran...@google.com wrote: The whole sln -- just got a new machine, got the tarball, gclient sync --force -- and trying to build chrome.sln. I was wondering if it could be interference between my cygwin, and the cygwin commands in the build? On Wed, May 20, 2009
