I installed Lazarus on my WinXP virtual machine from the fixes_30 branch a couple of weeks ago, but have not used it yet (Xmas and all). Now I just started and found that I needed to set the editor up for my usual look.
One thing that is very strange is the font size, when I set it to 10, like I usually do in Delphi, the text is nearly invisibly small!

Is Lazarus using a different size unit than Delphi is? I usually set the font to 8 or 10 to get a comfortable view of the code but that does not work at all in Lazarus, here I need to set it to 16 or even 18. What is the reason for this?

Bo Berglund wrote

Is Lazarus using a different size unit than Delphi is? I usually set the font to 8 or 10 to get a comfortable view of the code but that does not work at all in Lazarus, here I need to set it to 16 or even 18. What is the reason for this?

I had the same problem under my Linux system. I made a copy of my existing profile, to separate the two Lazarus versions. My fonts was very small in the editor. Then after I got some more updates my fonts were suddenly HUGE, and I had to set the font size back to normal again (11pt on my system).

That's Lazarus for you - after an update, you just don't know what you are going to get.

Graeme.

Not sure what to make out of this statement?

You upgraded to an Snapshot or SVN version. Not to a release. I know you have been around more than long enough to read the big fat statement about snapshots?

I believe a change in font size, is far less severe than "eat your disk"?

So if you had the courtesy to explain what exactly your above statement tries to tell us?
Did you just forget to mention you were talking about snapshots?

Because I am sure (but I do not have the time to prove it), if I search your fpgui git repository (argh now I said it GIT => aka "jehova" by brian) I could somewhere find a revision that had a bug or regression of some kind? How would you like if I was then telling everyone that your software is unreliable and all that? (just forgot to mention it wasn't a release...)

About the font issue, IIRC (!) there was a bug, and the 1st fix had side effects... well that happens

Oh, and about snapshot vs release => yes releases have bugs too (browse mantis). But again that is something that applies to every software in the universe (another unproven statement, but please find me a bug-free software if you can (one that does more than "hello world")

If you dislike the Lazarus project, or the way it is maintained, then that is fine. You are free to express this. Some of you post included valuable critics, and yes some of those never made a difference, for whatever reason....

But if your purpose is to distribute slander and libel, then all I can say is that it says a lot about your personality....

Have a nice day.

Regards
Martin

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20110115/bf6e415e/attachment-0001.html>

I've seen that as well over the last few weeks. It's also worth noting that different editors were affected differently, i.e. comparing the main editor with the one for SQL queries.

That's Lazarus for you - after an update, you just don't know what you are going to get.

No, that's a version control system for you.

What would be helpful would be a scoring system: if any of the developers knows that he's just checked something in that might result in an inconsistency a branch (or tag etc.) gets bumped from green to amber, or to red if he knows there's something dodgy that will be fixed in a few minutes.

I thought about something like this too. Plus svn trunk users could "vote for" revision they use if it is stable enough for them. These two measures combined would help users who want more stable IDE+LCL to jump between good revisions and avoid bad ones.

The side effect though is that errors would be found later than sooner, i.e. the number of bad revisions would theoretically increase when using this approach.

I suspect that developers would be unhappy about something like that, although that might be a gut rather than reasoned reaction.

Possibly the simplest thing would be if checkins wrote a timestamp to a file in the branch/tag's root, that would mean that a client script could hold off if it saw (for example) that there was an update less than five minutes ago just in case it's part of a flurry of development activity.

Hi Martin,

I simply forgot about the smiley face at the end of that sentence [though in the past, even with a smiley face some still interpret some of my messages as hostile]. I just can't win.

And yes, I know all to well that Trunk is by no means guaranteed to be stable. Hence I update my Lazarus very seldom these days. Unfortunately I had to update to 0.9.30 (fixes branch) to help bug-testing - as per Vincent's request in this mailing list.

Mark Morgan Lloyd wrote

What would be helpful would be a scoring system: if any of the developers knows that he's just checked something in that might result

Simply having 2 or 3 VM images will go a long way too... The developer simply needs to run his changes on the top two or three systems before committing to SubVersion (though with SubVersion not having local commits, that might be an issue in itself).

eg: For my development work, each VM (I have about 5 setup) use shared folders all pointing to my main development machines programming directory. So no need to copy code between VM sessions, just start them up. So whenever I make big scale changes, they get tested on those VM's before I push those commits to the public repository. For small and simple one line fixes you don't need such extensive testing (well normally).

But then, I have no say in how each Lazarus developer works. Each one for there own.

It aint a competition, but smileys I have myself forgotten before. Thou art forgiven ;)

So this turned out to be a bug in the 9_30 fixes branch...

By setting a different size number than I usually do the editor is working fine as far as I can see.
Do you know if the font issue is solved by now (on the fixes branch) so I can update from SVN and rebuild?

A regression? That is it wasn't in 0.9.28.2?
Or a normal bug? Like one of the many other bugs that have not yet been fixed, and will not be fixed for 0.9.30?

Has this bug ever been reported on mantis? In regards with Win-XP => there was an issue with gtk2

Also to clarify, you mean:
- the current Lazarus displays a font that is set to 10, differently than the released Lazarus 0.9.28(.2) did?
- or: Lazarus displays a font that is set to 10, differently than Delphi (or whatever your app of reference is)?

If it is the 2nd: A font has a height and a Size. Height=10 differs from Size=10.
I don't currently know, if in Lazarus you use the Height or Size, and I do not know what Delphi or other apps do...

There is however: http://bugs.freepascal.org/view.php?id=17694 (which is affecting win too). So depending how you set the font size, you may experience this bug.

In my case the desktop (Gnome) was a font size of 11 (I believe this to be points otherwise I would not be able to read anything. Lazarus IDE (everything but the source editor) displayed correctly. But the source editor has a very small text size. Using the font dialog I had to bump the font size to 16pt to display like the rest of my desktop and all other GTK apps.

Then after some date (or maybe I switched to 0.9.30 fixes) I had to reset the font size again, because then suddenly the source editor font was huge. I'll double check tomorrow at work my different source editor settings between Trunk and 0.9.30 Fixes branch.

And yes I know in Lazarus you can specify a font height as pixels (I think it's the negative font size value or something). I never do that, I use the standard GTK font dialog to specify my font size from the list.

It seems there are 2 or 3 issues being mixed up:

Bo Berglund mentioned XP. So that differs from your case.

1)
Yes, I do remember the gtk change, It changed my gtk font too. IIRC there was a 2nd fixed, after which you would get the normal font size again. (but I may recall that wrong).

In any case the gtk-case refers to the issue, where the editor font size suddenly changed *without* any user changes to the configuration. And I am quite sure that particular bit was looked into and corrected.

2)
Independent of this, there is
http://bugs.freepascal.org/view.php?id=17694 which exists on all platforms, which is *not* a regression (to the best of my knowledge this has always been like this).
This 2nd issue, and the fact that the Lazarus editor may take it's font size in height, while maybe every other app takes it as size, are independend of the first "sudden no user gtk only change of font size" issue.

Yes this 2nd bug is still present in 0.9.29 fixes, like many other bugs. It is not a secret that not all bugs can be fixed for a release. This bug belongs to the ones that will not be fixed for this release. Since it isn't a regression, I don't see why in the case of this particular bug, this should be any more of a problem, than with any other not-yet-fixed bug?

3)
As for Bo Berglund's issue, I do not know if he refers to the 2nd issue above, or a 3rd issue (but since he said XP, it wouldn't be the 1st)

Martin

Graeme Geldenhuys wrote

And yes I know in Lazarus you can specify a font height as pixels (I think it's the negative font size value or something). I never do that, I use the standard GTK font dialog to specify my font size from the list.

What I did was to start customize my brand new Lazarus and this involves changing the colors around quite a lot since I am used to a certain style since many years of Delphi work.
Anyway, I also wanted to reduce the size of the font so I went to the "Editor Font Height" box (which had defaulted to -17 if I remember correctly). Here I changed it to 10 expecting a font size similar to what I get when setting tha size in other Windows applications like Notepad etc. But it caused the very small size to appear. Finally I set it to 16 and it is OK.

I never noticed the sign thing until I read this reply, though. The default was negative....

Since people ask things like "which version are you using" can you please tell me how I can find out if I have downloaded from SVN and specified the fixes branch of 0.9.30?
I have no experience whatsoever with SVN, just installed it and used the suggested command on teh WIKI to get the files....

Martin wrote

As for Bo Berglund's issue, I do not know if he refers to the 2nd issue above, or a 3rd issue (but since he said XP, it wouldn't be the 1st)

Well, I just asked because I found it strange on my Windows XP Pro system that setting the editor font size to 10 would cause it to nearly disappear....
I did not count on this big debate being created. And I was later just believing that this was a bug in the IDE, that probably would be fixed. Being in the IDE and having a workaround does not make it a big issue at all for me, though. :)

I am happy just as it is while I am working on my first Lazarus program, a GUI interface for GPG that I eventually will have integrated into the Windows 7 Explorer, something no GPG app I have yet tested is able to do for some strange reason.

So you got the size versus height issue.
Which is indeed a very old bug (afaik), and which is unfortunately not fixed in the upcoming release.

In the help menu, abut lazarus should have the version.

If you donloaded from svn, you can lso check with the svn tool for the revision number.

Well, it didn't create it, just added a bit of confusion, because there is/was a similar 2nd issue, which actually is not related...

And also because from your mail ti was not clear (at least to me) if you just pointed out this bug, or if you where indicating a regression. That is, if your statement meant:
- you found that bug in fixes branch (but have no idea if it existed before);
- you found it in fixes, and by saying "in fixes" you meant it is a regression

And I was later just believing that this was a bug in the IDE, that probably would be fixed. Being in the IDE and having a workaround does not make it a big issue at all for me, though. :)

It will be fixed, eventually...

In my case this is what I read from the Help/About: Version #: 0.9.29
Date: 2011-01-02
SVN rev no: 28857

Strange that it says 0.9.29 since I retrieved the 0_9_30_fixes...

If you downloaded from svn, you can also check with the svn tool for the revision number.

I guess that is what the About box shows as 28857?

Bo Berglund wrote

In my case this is what I read from the Help/About: Version #: 0.9.29 Date: 2011-01-02 SVN rev no: 28857

You can right click inside the About dialog. The popup menu will copy all the details to the clipboard, ready to paste in a email. :)

Strange that it says 0.9.29 since I retrieved the 0_9_30_fixes...

I know. I reported it as a bug, but Vincent thinks it ok. I would have expected '0.9.30 rc1' as the version number.