Ralf Carl a ?crit :> Hello Erwan, > > i just want to inform you, that there is still no memory entry in the > main menu (hdt_034 & hdt_035). > Typing in DMI mode "show module 0" etc. the memory modules are listet.You mean the menu mode isn't working whereas the cli mode works ? Do you confirm this , IF yes, It seems the Type 6 (the only one you have as you lack of type 17) isn't present from the menu as the commit only add it to the CLI. http://git.zytor.com/?p=users/erwan/hdt.git;a=commitdiff;h=4899adbd7ad688fb6a6b8dc2f8aad54719168e09 So we'll have to fix that for 0.3.6.> > Since v 0.3.5 the graphic output is very slow. It was very fast in v > 0.3.4 and the older versions. > Did you change something for the graphic output ?Yes, some changes occured to use the libansi. What's your test case to feel this ? On which system are you working on ? I'm putting the list in CC to allow other people to follow this discussion.
On Thu, Nov 12, 2009 at 11:20:45AM +0100, Erwan Velu wrote:> > Since v 0.3.5 the graphic output is very slow. It was very fast in v > > 0.3.4 and the older versions. > > Did you change something for the graphic output ? > Yes, some changes occured to use the libansi. > What's your test case to feel this ? On which system are you working on ?If used over serial, it would be great to have a dump of the ansi sequences for debugging. Thanks, -- Pierre-Alexandre Meyer
Erwan Velu a ?crit :> Ralf Carl a ?crit : > >> Hello Erwan, >> >> i just want to inform you, that there is still no memory entry in the >> main menu (hdt_034 & hdt_035). >> Typing in DMI mode "show module 0" etc. the memory modules are listet. >>Can you make a try in the menu mode with the attached hdt version. This one is featuring a patch that should fix issues you found in the menu mode. If it works, I'll commit that. Thx. Erwan, -------------- next part -------------- A non-text attachment was scrubbed... Name: hdt.c32 Type: application/octet-stream Size: 172472 bytes Desc: not available URL: <http://www.zytor.com/pipermail/hdt/attachments/20091112/dd6599d8/attachment-0001.obj>
On Thu, Nov 12, 2009 at 10:36:02PM +0100, Ralf Carl wrote:> attached you will find some screenshots from my system. > There is also a dmidecode-output and a short lshw-output.Thank you for the detailed report. I have noticed quite a few bugs. For tracking purposes, the following tickets have been created on hdt-project.org: #22 Garbage in DMI type 6 output #23 Wrong bank connections information (DMI type 6) #24 Slow menu Could you provide more information on #24? Are you using Serial? What is your configuration file? Thanks, -- Pierre-Alexandre Meyer
Good evening Erwan, On Thu, Nov 12, 2009 at 07:58:18PM -0800, Pierre-Alexandre Meyer wrote:> Thank you for the detailed report. I have noticed quite a few bugs. For > tracking purposes, the following tickets have been created on > hdt-project.org: > > #22 Garbage in DMI type 6 output > #23 Wrong bank connections information (DMI type 6)I may have some fixes for these two bugs (unfortunately, I cannot test them). Could you review, get my changes and produce an official build for Ralf Carl to test (request-pull below)? For convenience, I have posted my build at https://www.hdt-project.org/wiki/hdt-0.3.6 Thanks, -- Pierre-Alexandre Meyer The following changes since commit ddf974948d8a28ef7f2090e7ab20d141f47268b0: Erwan Velu (1): hdt: Fixing subrelease are available in the git repository at: git://git.zytor.com/users/pam/hdt-pierre.git hdt-0.3.6 Pierre-Alexandre Meyer (2): dmi: fix DMI type 6 output dmi: fix bank connections information com32/gplinclude/dmi/dmi_memory.h | 6 +++--- com32/gpllib/dmi/dmi_memory.c | 10 +++++----- 2 files changed, 8 insertions(+), 8 deletions(-)
On Fri, Nov 13, 2009 at 08:56:02PM +0100, Ralf Carl wrote:> Did you change to another vesa mode in hdt-036 ? > Maybe the voodoo 3 graphic card has problems with some vesa modes ?hdt-0.3.5 features a rewrite of the UI. It doesn't use BIOS calls anymore but ANSI escape sequences (so that HDT can be used via serial). Anyone else on the list seeing a slowdown with hdt-0.3.5?> p.s. there is no memory entry in the menu in the very last version > 0.3.6.alpha.Is 0.3.6.alpha http://hdt-project.org/attachment/wiki/hdt-0.3.6/hdt.c32 or a custom build? Erwan: I couldn't find a related commit in your tree. Are you on it? Thanks, -- Pierre-Alexandre Meyer
On Sat, Nov 14, 2009 at 07:57:57PM +0100, Ralf Carl wrote:> p.s. have you seen the short avi file ?Yes, thank you for making it. It is really bad and definitively a blocker moving forward. I can't reproduce it here though. Could you try menu.c32 with 3.84-pre2 and a few (dummy) entries in your configuration file? Try going down and up to see if you experience the same delay. By the way, how did you test the 0.3.6 build (which version of Syslinux core)? Thanks, -- Pierre-Alexandre Meyer
On Sun, Nov 15, 2009 at 10:22:30PM +0100, Ralf Carl wrote:> after many hours of testing i removed the line "SERIAL 0" in the > syslinux.cfg file. > > And now - all the speed problems with the menu of hdt are gone. > Everything is fast and smooth ! > > If the line "SERIAL 0" remains in the config file, > the module menu.c32 is very slow when scrolling the menu.Great news! HDT menu and menu.c32 now share the same codebase to render graphics (ANSI console). Testing with the latter is useful to check if a bug is in HDT or in Syslinux libs. I have closed https://www.hdt-project.org/ticket/24.> All the testing was done with: > > syslinux 3.82 , 3.83 , 3.84pre2 > hdt 0.3.4 , 0.3.5 , 0.3.6 alphaFor the future, it is always useful to add in a bug report which version of Syslinux core you are using. When running HDT, you are really touching two codebases, hdt.c32 and Syslinux core. This is the reason why we always specify on the website which version of Syslinux you should be using with the build we provide. Enjoy HDT! Thanks, -- Pierre-Alexandre Meyer
Ralf Carl a ?crit :> Hi Erwan, > > i have done a new test with hdt_036alpha (16.11.2009) and syslinux 3.83 > and YES :-) there is a memory entry in the main menu - great ! > > There is only one thing remaining: > switching from menu to cli mode and typing "show" > there is a message "No memory module found" >Good point. I see what to do for fixing this. I'll manage to make a patch asap. Thanks for reporting this bugs.
Ralf Carl a ?crit :> Hi Erwan, > > how does it work ? > the total amount of memory should be 512 MB !We rely on that detection on what the dmi table says. In our case, we are coherent with your DMI. I just try to understand why it says that.> > > Menu summary: > > - i think "mem bank 00" etc. is missing in the summary.Yeah, good point, fixed.> - field CPU Model is empty.Your cpu doesn't support identification. I've been trying to workaround that. Tell me if it works.> - graphic fault at the border line of the summary >I didn't understood what you mean. Please download https://www.hdt-project.org/attachment/wiki/hdt-0.3.6/hdt.c32 MD5 is : 9c7494cbecf134e04d13912a72c52fee
On 11/15/2009 02:48 PM, Pierre-Alexandre Meyer wrote:> On Sun, Nov 15, 2009 at 10:22:30PM +0100, Ralf Carl wrote: >> after many hours of testing i removed the line "SERIAL 0" in the >> syslinux.cfg file. >> >> And now - all the speed problems with the menu of hdt are gone. >> Everything is fast and smooth ! >> >> If the line "SERIAL 0" remains in the config file, >> the module menu.c32 is very slow when scrolling the menu. > > Great news! HDT menu and menu.c32 now share the same codebase to render > graphics (ANSI console). Testing with the latter is useful to check if > a bug is in HDT or in Syslinux libs. >Incidentally, with this change you should be able to use the VESA screen for HDT if you want to. -hpa
On Wed, Nov 18, 2009 at 08:00:06PM +0100, Ralf Carl wrote:> I have another one: > > entering to dmi mode and just pushing the TAB key, a list of available commands is shown from "show" to "memory". > Then the program freeze for about 1 minute. > A line with unreadable chars is displayed, then the output "clear" to "reboot" is displayed.I may have a fix in my hdt-0.3.6 branch. Could you please pull? Thanks, -- Pierre-Alexandre Meyer The following changes since commit 787b25c5d9bf8197aa00bcb42e3615343cde6253: Erwan Velu (1): hdt: Using dmi when cpu doesn't support identify are available in the git repository at: git://git.zytor.com/users/pam/hdt-pierre.git hdt-0.3.6 Pierre-Alexandre Meyer (4): hdt: fix memory corruption in CLI hdt: more memory corruption fixes hdt: trivial printf fix (CLI) hdt: fix double lines in the menu com32/hdt/hdt-cli-disk.c | 1 - com32/hdt/hdt-cli.c | 51 +++++++++++++++++++++++++++-------------- com32/hdt/hdt-menu-summary.c | 19 ++++++++------- 3 files changed, 43 insertions(+), 28 deletions(-)
Hey Ralf, Please found at this url https://www.hdt-project.org/attachment/wiki/hdt-0.3.6/hdt.c32 a new version that should fixes all the troubles you reported. Pierre and I did commit some fixes for your bugs. Can you confirm that it works better for you too ? Sincerly, Erwan