stgatilov

I knew such complicated changes won't go unnoticed @Gadavre has run into a bug caused by this change: I uploaded a new version of tdm_installer with a fix.

Yes, I guess it can be reproduce in several ways. I have fixed the bug and pushed update to the server. Could you please try again and report if the problem still happens? I booted up local mirror and did something like 15 random hops between TDM versions, no fails.

I can reproduce this issue when updating from 2.09b to 2.07. It look very worrisome to me. I'll investigate. There are a lot of simpler distribution approaches as long as players don't want differential updates. The installer does that internally. Also, checksum works perfectly for single file, but for many files it would be a headache.

Do you have Windows 7? Does it have any kind of service pack installed? Can you go to C:\Windows\System32, find dbghelp.dll there, Right-click it in Explorer and look properties. You can see its version there in "detailed" tab or something like that.

I think one day I'll get member offsets using libclang. And I'm pretty sure typeinfo will be removed soon... not even sure which will happen earlier.

How did you get away without #define private public ? As far as I remember, it is required to get member pointers.

Hello, @Flaniel! Some random ideas: You have some kind of antivirus, which does not like the executables. I'd suggest looking through antivirus logs to see if it has ever messed with TDM files. You have hardware problem. It might be HDD/SSD, or RAM. Better check them to be sure. I can probably say something else if you provide the logs of tdm_installer: they are called like tdm_installer_1637476178.log. I'll be able to see exact error message at least. If only executables disappear, you can easily restore them by unpacking tdm_shared_stuff.zip.

Find or get tdm_installer. Use instructions in beta testing thread to get beta210-01.

Ha... the outline does not show at all on some doors,which is also not fixable. Unless everyone likes the depth-ignoring outline, but we have already left this behind I hope.

I have finally committed all the updates to the mission database. The mirrors will catch up soon. I jumped the gun a bit, since originally I wanted to commit them closer to the end of beta. But it is hard to beta-test 2.10 with all the old missions around...

Mission "Away 0: Stolen Heart" @Geep and @_Atti_. It overrides mainmenu_background.gui for a custom background. I deleted this file, instead added mainmenu_background_custom.gui with following contents: windowDef BackgroundCustomAway0 { rect 0, 0, 640, 480 background "guis/assets/custom_title" BACKGROUND_DEFAULT_BEHAVIOR(BackgroundCustomAway0) } Also added the following to mainmenu_custom_defs.gui: //stgatilov #5323: set custom background to main menu: #define MM_BACKGROUNDS_META_ALLSIMPLE BackgroundCustomAway0 Tested the changes: in 2.10 it works fine, background is modified successfully. In 2.09 everything is fine, but default background is shown instead of the custom one.

Since a major update (v4) came after I did that, my old pk4 archive is no longer relevant. So I took fresh pk4 from mission SVN, applies exactly the same changes, and committed a new pk4. Checked that it works well on 2.10 beta.

There was update to Briarwood Manor on October, 11-th, hence the file which I prepared is no longer relevant. I took the latest version from missions SVN, applied the same changes that I described in quoted post, and committed to missions SVN.

It turned out that the path "guis/assets/dragon_claw_background" was wrong. Changed it to "guis/assets/mainmenu/dragon_claw_background".

I think you need to set "r_newFrob 0" and "r_frobOutline 0" to restore old behavior. I put asterisk in "r_frob*" to denote that there are many cvars with this prefix. Game console supports autocomplete, i.e. you can write "r_frob", then press Tab, and see all cvars/cmds with this prefix. Then you can execute each one individually and read description. I recall autocomplete did not work on Linux build though, but maybe recent GLFW changes fixed that?... Also, you can type "listCvars" to see all cvars, and maybe "listCvars r_frob*" or "listCVars r_frob" will show you the list of frob-related cvars.

Why no? It looks like it helped, but other packages are missing too. Do you have bincrafters remote? As far as I remember, it is not enabled by default on fresh installation, you need to add it manually.

Yes, padding is a problem. One possible thing is to check every 8 bytes instead of 4 bytes: I guess it is already done like that in 64-bit builds. Of course, it allows the check to miss uninitialized members, but I don't think there is a way around it.

Beta releases will be relevant for 2 months, and dev releases for 10 months after that. Are you sure you won't ask to return back the old order in February/March?

Anyway, the generated typeinfo is wrong. If you try to compile TDM with it, you'll get a lot of errors. Most of the are due to not understanding C++11, and it is not wise to eradicate C++11 from headers or trying to #ifdef every little thing.

By #ifdef-ing by ID_TYPEINFO. I have committed a fixed version. It turned out your problem was mainly due to ID_TYPEINFO being defined when building TypeInfo C++ code. I committed fixed version.

Yes, except for silencing #error directive.

Does it help if you replace "zlib/1.2.11@conan/stable" with "zlib/1.2.11" in conanfile.py ?

I did not consider it important enough In fact, the announcement is for experienced players, and they will immediately notice this change, but they are probably OK with either approach. It is the new players who will benefit most (much less confusion), but I suppose they won't read the beta announcement anyway.

I lock this thread, because beta testing has started. There will be no dev builds until 2.10 is released. After that I'll extract all the comments to a separate unpinned thread, to start afresh for a new dev. cycle.

No, I committed all the changes to SVN at that time. But typeinfo parser is very fragile. I'm sure it quickly broke after that due to normal development.