October 2014 - Wine-devel - list.winehq.org

Re: [PATCH] vbscript: Implemented String
by Marvin 19 Oct '14

19 Oct '14

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at https://testbot.winehq.org/JobDetails.pl?Key=9624 Your paranoid android. === w2000pro (32 bit) === No test summary line found === wxppro (32 bit) === No test summary line found === wvistau64 (32 bit) === No test summary line found === w2008s64 (32 bit) === No test summary line found === w7u (32 bit) === No test summary line found === w8 (32 bit) === No test summary line found === wvistau64 (64 bit) === No test summary line found === w2008s64 (64 bit) === No test summary line found

1 0

Re: comctl32: Use strcmpW instead of pointer comparison to ensure treeview updates correctly
by Nikolay Sivov 19 Oct '14

19 Oct '14

On 10/19/2014 18:44, Dan Bassi wrote: > + /* Old text is a null pointer due to HeapAlloc() failing, therefore unable to perform strcmpW so assume text has changed */ > + if (!tiOld->pszText) > + return TRUE; > + I don't think you should necessary assume it's changed. > /* Text has changed and it's not a callback */ > - if ((tvChange->mask & TVIF_TEXT) && (tiOld->pszText != tiNew->pszText) && > + if ((tvChange->mask & TVIF_TEXT) && (strcmpW(tiOld->pszText, tiNew->pszText) != 0) && > tiNew->pszText != LPSTR_TEXTCALLBACKW) > return TRUE; This is wrong, you can't compare if it's a special callback value. > + originalItem.pszText = HeapAlloc(GetProcessHeap(), 0, originalItem.cchTextMax * sizeof(WCHAR)); Normally comctl32 code uses Alloc()/Free() functions. This patch needs tests first, those tests should cover all possible combinations regarding callback cases, different pointers but same data, same data but different pointers, maybe something else.

1 0

Re: wined3d: Correctly display fog for right-handed projection matrix (try 7)
by Stefan Dösinger 18 Oct '14

18 Oct '14

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, This one looks good to me. Am 2014-10-12 17:08, schrieb Joachim Priesner: > + enum > + { > + COLOR_ALPHA_0x00 = 0x00000000, > + COLOR_CLEAR = 0xffff00ff, > + COLOR_FOGGED = 0x0000ff00, > + COLOR_UNFOGGED = 0x00ff0000 > + }; > + > + /* Fill the null-shader entry with the FVF (SetVertexShader is "overloaded" on d3d8). */ > + DWORD vertex_shader[3] = {D3DFVF_XYZ | D3DFVF_DIFFUSE | D3DFVF_SPECULAR, 0, 0}; Do you need the specular color on Windows as well, or to make this work right on Wine? I suspect that the NONE/NONE test doesn't work right on Wine because we do not set a controlled fog coordinate in wined3d/state.c, load_vertex_data(), in the case where WINED3D_FFP_SPECULAR is absent and similarly in drawprim.c, drawStridedSlow(). In either case that's no reason to hold up this patch any longer IMHO. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQIcBAEBAgAGBQJUOp0lAAoJEN0/YqbEcdMwgD4P/iLeSfFbE/jEsJ+Yux1lR9PC FDoJ50dxiedieOJWl4TbvIgr5uNyXnBm9YtPiiLgZJNfwoJ2znYjYDsA4h2yZ8FU lrgHwWcAo1ZvlPS0QolSYfx7QH4TpGIKklfEr7Qe4LQgftJjHNGioDuK1zpT+E3e JjcZOXDKjOP4ACNu5awEBq9v427TxOT9E+GJQWrq/cxbCYRbDurFwuFS2Jb1o/j+ 9zEoy8qiaQ+6v0j8+RcAkBTMdg6DBwSOYsH7P/KjrnzUkj+yH1w+R9dcnHMdaeMP RRoJSLsGzfANXF05vn7vE5Q5QzOjhXtIWWL/yun/IAL/EofZrXyVikoGP28wxUV1 QBICfDUvel4dhnYxNzNGUMQXC7gaXmJI3q12gemhkX4vj2W/FxVIaSN7vr5j+5rp 2ijdi9rexyBopbDM/9ay599GSNjXrTljkv8sEH23JMIpKR2BP7afyih6Hta9e/zf Ao9LtFwyBye2F9HTyM9G10MBrnVdNIKhJ9ZY30bRhomtXrfP1HuRvvpfRPja9gjZ M/aiwqPM6a3vfdJ/kIOra72vNZAoE9ccEOuDPQqMhg2ccoYvLUg6+LGvKRLanmOO lCsKiFYUChwJlebbo98+bfhMiR7TBI5fkwXOxR4coDu9jIl1oP2GblaL5Sul+zg1 rXbnkL2Pltlb/ZaW33l0 =Ozzv -----END PGP SIGNATURE-----

3 6

Re: [PATCH 1/3] kernel32: Support UTF-7 in MultiByteToWideChar. (try 3)
by Michael Stefaniuc 17 Oct '14

17 Oct '14

Hello Alex, I'll have a go at reviewing your patches as Alexandre didn't take them as is. On 10/13/2014 05:54 AM, Alex Henrie wrote: > --- a/dlls/kernel32/locale.c > +++ b/dlls/kernel32/locale.c I'm pulling this diff chunk to the front as it is wrong and some of your code is over-complicated based on that assumption. > @@ -1963,7 +2119,7 @@ BOOL WINAPI EnumSystemCodePagesW( CODEPAGE_ENUMPROCW lpfnCodePageEnum, DWORD fla > * flags [I] Character mapping flags > * src [I] Source string buffer > * srclen [I] Length of src (in bytes), or -1 if src is NUL terminated > - * dst [O] Destination buffer > + * dst [O] Destination buffer, or NULL to compute the required length > * dstlen [I] Length of dst (in WCHARs), or 0 to compute the required length The only case where dst == NULL is when dstlen == 0 but that already implies computing the required length. dst == NULL and dstlen != 0 is an error and MultiByteToWideChar() bails out. > @@ -1954,6 +1954,162 @@ BOOL WINAPI EnumSystemCodePagesW( CODEPAGE_ENUMPROCW lpfnCodePageEnum, DWORD fla > > > /*********************************************************************** > + * write_to_w_string > + * > + * Helper for utf7_mbstowcs > + * > + * RETURNS > + * TRUE on success, FALSE on error > + */ > +static inline BOOL write_to_w_string(WCHAR *dst, int dstlen, int *index, WCHAR character) This isn't a good name for the helper as that's not what it does. It is more a "count and optionally write". Also prefixing it utf7 would make it clear that is just a helper for utf7_mbstowcs and not a generic function and thus would remove the need for the comment that it is just a helper for utf7_mbstowcs. I know you're using this helper to avoid code duplication between the length calculation and the conversion unlike wine_utf8_mbstowcs(). But it complicates the handling of available space in dst. Without seeing the code for both approaches I'm not sure what is best so I'll review it as is. > +{ > + if (dst) You have to force dst = NULL in utf7_mbstowcs() for this to work. Just checking for dstlen > 0 should avoid that. > + { > + if (*index >= dstlen) > + return FALSE; > + > + dst[*index] = character; > + } > + > + (*index)++; > + > + return TRUE; > +} > + > +/*********************************************************************** > + * utf7_mbstowcs > + * > + * UTF-7 to UTF-16 string conversion, helper for MultiByteToWideChar > + * > + * RETURNS > + * On success, the number of characters written > + * On dst buffer overflow, -1 What about the case where it fails to decode a character? E.g. wine_utf8_mbstowcs() has -2 for that case which then translates to SetLastError(ERROR_NO_UNICODE_TRANSLATION). I looked at the tests but you didn't seem to test the GetLastError in the cases with the invalid input characters. > + */ > +static int utf7_mbstowcs(const char *src, int srclen, WCHAR *dst, int dstlen) > +{ > + static const WCHAR base64_decoding_table[] = { Why is this table an array of WCHAR? The entries aren't WCHARs. > + /* \0 */ -1, /* \x01 */ -1, /* \x02 */ -1, /* \x03 */ -1, This table is really hard to read even in a editor that does syntax highlighting. As you just need 4 characters per column you could have sixteen columns and merge the comments to one comment at the end of the line. Something like this would be more readable as well as being more compact: -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, /* 0x00 - 0x15 */ .... /* '0' - '?' */ > + /* \x04 */ -1, /* \x05 */ -1, /* \x06 */ -1, /* \a */ -1, > + /* \b */ -1, /* \t */ -1, /* \n */ -1, /* \v */ -1, > + /* \f */ -1, /* \r */ -1, /* \x0E */ -1, /* \x0F */ -1, > + /* \x10 */ -1, /* \x11 */ -1, /* \x12 */ -1, /* \x13 */ -1, > + /* \x14 */ -1, /* \x15 */ -1, /* \x16 */ -1, /* \x17 */ -1, > + /* \x18 */ -1, /* \x19 */ -1, /* \x1A */ -1, /* \e */ -1, > + /* \x1C */ -1, /* \x1D */ -1, /* \x1E */ -1, /* \x1F */ -1, > + /* */ -1, /* ! */ -1, /* " */ -1, /* # */ -1, > + /* $ */ -1, /* % */ -1, /* & */ -1, /* ' */ -1, > + /* ( */ -1, /* ) */ -1, /* * */ -1, /* + */ 62, > + /* , */ -1, /* - */ -1, /* . */ -1, /* / */ 63, > + /* 0 */ 52, /* 1 */ 53, /* 2 */ 54, /* 3 */ 55, > + /* 4 */ 56, /* 5 */ 57, /* 6 */ 58, /* 7 */ 59, > + /* 8 */ 60, /* 9 */ 61, /* : */ -1, /* ; */ -1, > + /* < */ -1, /* = */ -1, /* > */ -1, /* ? */ -1, > + /* @ */ -1, /* A */ 0, /* B */ 1, /* C */ 2, > + /* D */ 3, /* E */ 4, /* F */ 5, /* G */ 6, > + /* H */ 7, /* I */ 8, /* J */ 9, /* K */ 10, > + /* L */ 11, /* M */ 12, /* N */ 13, /* O */ 14, > + /* P */ 15, /* Q */ 16, /* R */ 17, /* S */ 18, > + /* T */ 19, /* U */ 20, /* V */ 21, /* W */ 22, > + /* X */ 23, /* Y */ 24, /* Z */ 25, /* [ */ -1, > + /* \ */ -1, /* ] */ -1, /* ^ */ -1, /* _ */ -1, > + /* ` */ -1, /* a */ 26, /* b */ 27, /* c */ 28, > + /* d */ 29, /* e */ 30, /* f */ 31, /* g */ 32, > + /* h */ 33, /* i */ 34, /* j */ 35, /* k */ 36, > + /* l */ 37, /* m */ 38, /* n */ 39, /* o */ 40, > + /* p */ 41, /* q */ 42, /* r */ 43, /* s */ 44, > + /* t */ 45, /* u */ 46, /* v */ 47, /* w */ 48, > + /* x */ 49, /* y */ 50, /* z */ 51, /* { */ -1, > + /* | */ -1, /* } */ -1, /* ~ */ -1, /* \x7F */ -1 > + }; > + > + const char *source_end = src + srclen; > + int dest_index = 0; > + > + DWORD byte_pair = 0; > + short offset = 0; > + > + /* MultiByteToWideChar guarantees that srclen > 0 */ > + assert(srclen > 0); This assert isn't needed. Just rewriting the first do..while loop to a while loop avoids it. That eliminates 3 lines of code of which one is a comment pointing to an unneeded assert. > + > + if (!dstlen) > + dst = NULL; Unneeded if the helper just checks for dstlen > 0. > + > + do Should be just "while (src < source_end)" to avoid the assert from above. > + { > + if (*src == '+') > + { > + src++; /* skip the + sign */ > + if (src >= source_end) > + break; > + > + if (*src == '-') > + { > + /* just a plus sign escaped as +- */ > + if (!write_to_w_string(dst, dstlen, &dest_index, '+')) > + return -1; > + src++; > + continue; > + } > + > + do > + { > + WCHAR sextet = *src; Same as for base64_decoding_table, sextet is not a WCHAR. > + if (sextet == '-') > + { > + /* skip over the dash and end base64 decoding */ > + /* the current, unfinished byte pair is discarded */ > + src++; > + offset = 0; > + break; > + } > + else if (sextet <= 127) > + { > + sextet = base64_decoding_table[sextet]; > + if (sextet == (WCHAR)-1) Unneeded cast. Using a signed type for sextet also avoids the -Wsign-compare compiler warning. Alternatively use ~0 instead of -1 if you want to keep an unsigned type. > + { > + /* -1 means that the next character of src is not part of a base64 sequence */ This upper part of the comment is better placed in the declaration of base64_decoding_table. > + /* in other words, all sextets in this base64 sequence have been processed */ > + /* the current, unfinished byte pair is discarded */ > + offset = 0; > + break; > + } > + } > + else > + { > + /* the next character of src is > 127 and therefore not part of a base64 sequence */ > + /* the current, unfinished byte pair is NOT discarded in this case */ > + /* this is probably a bug in Windows */ > + break; > + } > + > + byte_pair = (byte_pair << 6) | sextet; > + offset += 6; > + > + if (offset >= 16) > + { > + /* this byte pair is done */ > + if (!write_to_w_string(dst, dstlen, &dest_index, (byte_pair >> (offset - 16)) & 0xFFFF)) > + return -1; > + offset -= 16; > + } > + > + /* this sextet is done */ > + src++; > + } while (src < source_end); > + } > + else > + { > + /* we have to convert to unsigned char in case *src > 127 */ > + if (!write_to_w_string(dst, dstlen, &dest_index, (unsigned char)*src)) > + return -1; > + src++; > + } > + } while (src < source_end); > + > + return dest_index; > +} > + > +/*********************************************************************** > * MultiByteToWideChar (KERNEL32.@) > * > * Convert a multibyte character string into a Unicode string. bye michael

3 7

reg.exe platform specific test?
by Jonathan Vollebregt 17 Oct '14

17 Oct '14

In the reg.exe tests around line 181: > verify_reg(hkey, "dword8", REG_DWORD, &dword, sizeof(dword), > (sizeof(long) > sizeof(DWORD)) ? 0 : TODO_REG_DATA); What's the reason for this? Is there some platform specific quirk I should be aware of?

4 4

Re: Italian translation of Wine
by Matteo Bruni 17 Oct '14

17 Oct '14

2014-10-16 11:19 GMT+02:00 Davide Pizzetti <erpizzo(a)alice.it>: > This is an updated patch for the Italian translation of Wine. > Hi Davide, Please bear with me, but: @@ -30,7 +34,6 @@ msgid "&Install..." msgstr "&Installa..." #: appwiz.rc:65 -#, fuzzy #| msgid "" #| "The following software can be automatically removed. To remove a program " #| "or to modify its installed components, select it from the list and click " I still think you should remove those "#| msgid" lines when removing the fuzzy tags. If the program you're using to update the .po file doesn't help you with that, you might have to do it by hand by opening the .po file in a plain text editor. @@ -1682,11 +1682,11 @@ msgstr "Autenticazione del client" #: crypt32.rc:140 msgid "Code Signing" -msgstr "Firma codice" +msgstr "Firma del codice" #: crypt32.rc:141 msgid "Secure Email" -msgstr "Email sicura" +msgstr "E-mail sicura" #: crypt32.rc:142 msgid "Time Stamping" I don't like those changes, they seem mostly gratuitous to me (there are others similar instances below). If you're changing the "email" translation, make sure to be consistent. In general I'd avoid adding more stuff to this patch on each iteration, first get this patch in and then you can work on another one with more string translations. @@ -1834,7 +1834,7 @@ msgstr "Altro nome=" #: crypt32.rc:186 msgid "Email Address=" -msgstr "Indirizzo Email=" +msgstr "Indirizzo mail=" This looks wrong. @@ -2861,7 +2855,7 @@ msgstr "" #: cryptui.rc:127 msgid "Protects e-mail messages" -msgstr "Protegge messaggi e-mail" +msgstr "Protegge i messaggi mail" Same. @@ -4800,7 +4785,7 @@ msgstr "Nessuno spegnimento in corso.\n" #: winerror.mc:1381 msgid "I/O device error.\n" -msgstr "Errore I/O dell'unità.\n" +msgstr "Errore I/O del dispositiov.\n" #: winerror.mc:1386 This is still wrong. #: netstat.rc:48 msgid "Reset Connections" -msgstr "Connessione LAN" +msgstr "Resetta le connessioni" This means "Connessioni resettate". You can use "chiuse" instead, or probably even "annullate" as you wrote in another message translation. @@ -13103,7 +13080,7 @@ msgstr "Wine" #: wineboot.rc:46 msgid "The Wine configuration in %s is being updated, please wait..." msgstr "" -"La configurazione di Wine in %s è in aggiornamento, aspetta un momento..." +"La configurazione di Wine in %s sta venendo aggiornata, aspetta un momento..." I prefer the previous translation.

3 4

Re: server: Fix invalid memory access when copying KeyNameInformation data.
by Austin English 17 Oct '14

17 Oct '14

On Thu, Oct 16, 2014 at 11:13 PM, Sebastian Lackner <sebastian(a)fds-team.de> wrote: > Fixes https://bugs.winehq.org/show_bug.cgi?id=37411 for me. > > --- > server/registry.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > Works here as well, thanks for the fix. -- -Austin

2 1

todo_wine under the hood
by Jonathan Vollebregt 17 Oct '14

17 Oct '14

I was looking at verify_reg_() in programs/reg/tests/reg.c:61 and noticed some simple if statements: > if (todo & TODO_REG_TYPE) > todo_wine lok(type == exp_type, "got wrong type %d, expected %d\n", type, exp_type); > else > lok(type == exp_type, "got wrong type %d, expected %d\n", type, exp_type); > if (todo & TODO_REG_SIZE) > todo_wine lok(size == exp_size, "got wrong size %d, expected %d\n", size, exp_size); > else > lok(size == exp_size, "got wrong size %d, expected %d\n", size, exp_size); > if (todo & TODO_REG_DATA) > todo_wine lok(memcmp(data, exp_data, size) == 0, "got wrong data\n"); > else > lok(memcmp(data, exp_data, size) == 0, "got wrong data\n"); Simple enough, except if I pass TODO_REG_TYPE to this function it never even tries the size or data blocks. While this makes sense to implement (You can't test the data if you don't know it's size) I have no idea *how* it does it. I've tried the obvious - the mask values don't overlap in a way that would cause it to fallthrough. The lok and todo_wine macros don't have anything in them that returns. In fact the todo_wine macro appears to start a loop for some reason and all in all I'm very confused as to how exactly it's doing this. I'm curious as to how this works under the hood. I also need to stick a broken() in there somewhere but every time I do it falls through, and I don't want to rewrite the entire verify_reg_() function for a single exception. (If you're interested, reg add a REG_BINARY with an odd number of characters prefixes a 0 nibble, but suffixes it in XP)

2 4

Re: [PATCH 3/5] d3d9/tests: Add style and size tests for D3DCREATE_NOWINDOWCHANGES.
by Henri Verbeet 17 Oct '14

17 Oct '14

On 16 October 2014 17:07, Stefan Dösinger <stefan(a)codeweavers.com> wrote: > diff --git a/dlls/d3d9/tests/d3d9ex.c b/dlls/d3d9/tests/d3d9ex.c > index 11fe091..34b967a 100644 > --- a/dlls/d3d9/tests/d3d9ex.c > +++ b/dlls/d3d9/tests/d3d9ex.c > @@ -37,6 +37,7 @@ struct device_desc > unsigned int width; > unsigned int height; > BOOL windowed; > + DWORD behavior_flags; > }; > I don't like this much, I'd probably prefer "windowed" and "behavior_flags" to be merged into a single flags field and then have flags along the lines of e.g. CREATE_DEVICE_FULLSCREEN, CREATE_DEVICE_NOWINDOWCHANGES, CREATE_DEVICE_SWVP_ONLY, etc.

2 1

Re: server: KeyNameInformation returns the full name of the key. (try 3)
by Austin English 16 Oct '14

16 Oct '14

On Mon, Oct 13, 2014 at 10:25 AM, Akihiro Sagawa <sagawa.aki(a)gmail.com> wrote: > Try 3: > - Avoid allocating memory. (Based on Alexandre's comment, thanks.) > - Use k != root_key instead of k->namelen != 0. > - Add insufficient buffer test. > > --- > dlls/ntdll/tests/reg.c | 19 ++++++++++++++----- > server/registry.c | 25 ++++++++++++++++++++++++- > 2 files changed, 38 insertions(+), 6 deletions(-) > Hi Akihiro, This change breaks the tests on my machine (causing wineserver to crash and the ntdll/reg test to fail), see https://bugs.winehq.org/show_bug.cgi?id=37411 Could you please take a look? Thanks, -Austin

1 1