Newbie
Registriert seit: 19.06.2003
Beiträge: 11
| Problem: eMule 0.29a SlugFiller 0.33a
Downloads:
Version 0.33a - eMule 0.29a
Binary:
ed2k://|file|eMule-29a[SF-VQB]-v0.33a.exe|1074733|C82D5BCC6D07C33C528CABEC357C90C 0|/|sources,12.203.22.14:8754|/
Source:
ed2k://|file|eMule-29a[SF-VQB]-v0.33a-src.exe|1281460|FD77B67F78A7E0406881D8C7C92AD536|/|sources,12.203.22.14:8754|/
Version 0.33 - eMule 0.29a
Binary:
ed2k://|file|eMule-29a[SF-VQB]-v0.33.exe|1180062|6B4B4EEE4E911BC27732907CF3972B40 |/|sources,12.203.22.14:8754|/
Source:
ed2k://|file|eMule-29a[SF-VQB]-v0.33-src.exe|1660576|B276349BDFB357DF9B4587AAC84908B3|/|sources,12.203.22.14:8754|/
Features:
[1] Check diskspace v1.2(Tag: "SLUGFILLER: checkDiskspace"):
* Verifies all non-paused downloads have enough room in the temp partition to finish.
* If there isn't enough room, downloads are paused so that the remaining downloads do have enough room. Files with lower priority are more likely to be paused in favor of higher priority files. Files paused this way have a status of "Insufficient diskspace".
* Whenever a download is paused, resumed, stopped, or has it's priority changed, the free disk space is rechecked to see if previously paused files can be resumed, or new ones paused because of the current non-manually-paused downloads changing.
* v1.1: Rechecks the free diskspace every 15 minutes, just in case the free disk space on the temp partition has changed(files copyed, created, deleted, etc).
* v1.2: Can be disabled through the new settings page.
[2] Share with subdirectories v2.3(Tag: "SLUGFILLER: shareSubdir"):
* When a directory is shared, all of it's subdirectories are shared as well.
* Only the root directory is saved in the dat files, the subdirectories are rescanned every time the shared files are reloaded or refreshed. As a result, the dat file is smaller than one where all the subdirectories are shared manually, and if a shared directory's subdirectories change(renamed, created, deleted, etc), the change will be detected the next time the shared files are refreshed.
* v2: To share a directory with subdirectories hold the ctrl key, otherwise the directory is shared without it's subdirectories.
* v2: Directories shared with subdirectories are now saved in a seperate dat file. The subdirectories themselves are still not saved.
* v2: Once a directory is shared with subdirectories you cannot unshare/reshare any of it's subdirectories, and the checkbox next to them disappears.
* v2: You can now share files from network shares without having to map them to a drive first.
* v2.1: Local drives appear before network drives.
* v2.2: Network drives load in another thread.
* v2.3: The incoming directory is shared without subdirectories by default and marking it causes it to be shared with subdirectories.
[3] Transfer whole chunk v2(Tag: "VQB: fullChunk"):
* Upload slot cycling is set to cycle only after 9.28MB(the size of a complete chunk) were sent, so that anyone downloading will get at least one chunk complete and be able to hash it and reshare.
* v1.1: A downloader getting a slot with the small-file boost will only download according to the size of the file used for the boost, and not a complete chunk.
* v1.1: No longer counts compressed packets according to their compressed size.
* v2: You can now choose if to always upload 9.28MB, or to simply upload only from one chunk, cycling when the user tries to start downloading the next chunk.
[4] Push small files v1.1(Tag: "VQB: SFpush"):
* Files under 1 chunk get a priority boost in inverse ratio to their size. A 1 chunk file(or above) gets no boost, and a 9.28k(1/1000 chunk) file(and below) gets the maximal boost of times 1000.
* v1.1: Small file boost can be seen in an extra column in the upload window.
[5] Release slots(Tag: "SLUGFILLER: slotRelease"):
* You can set complete files to be release files. As long as there isn't a release file currently uploading, any request for a release file is auto-accepted.
[6] General Bugfixes and Tweaks:
* Known.met is saved after every upload is complete to preserve statistics in case of crash(Tag: "VQB: saveKnownMet").
* Reasks for files now have an added random delay of up to 5 minutes to prevent connection peaks caused by all reasks occuring all at the same time(Tag: "SLUGFILLER: spreadReask").
* Fixed source exchanged sending IDs instead of IPs. Now only sources with a valid IP are sent by IP(Tag: "SLUGFILLER: xchgByIP").
* Multi-layer sorting. If you sort a list by one column, then by another, the order of the sorting will be remembered, and when the items change it will fit the same order of sorting instead of just the last(Tag: "SLUGFILLER: multiSort").
* The statistics of duplicate known files are now merged and the statistics of part files are now saved in the known.met(Tag: "SLUGFILLER: mergeKnown").
* Recheck needed parts when known needed parts finished downloading patch, by Ottavio84(Tag: "enkeyDEV(Ottavio84):noNeededReask").
* Revised code for creating upload packets to one that doesn't need to assume your fragment size and doesn't time out for low slot bandwidth(Tag: "SLUGFILLER: noRefrag").
* Revamped multiple aspects of downloadlist sorting(Tag: "SLUGFILLER: DLsortFix").
* Downloads can take sources from search results and browsed user files(Tag: "SLUGFILLER: searchCatch").
* Download links can contain hostname sources, instead of just ip(Tag: "SLUGFILLER: hostnameSources").
* If multiple files are selected while entering a new comment, the new comment is applied to all of them(Tag: "SLUGFILLER: batchComment").
* Added checks to startup timer, so that if the controls aren't yet initialized, even after the delay, it waits for another timer cycle until they are before trying to initialize(Tag: "SLUGFILLER: doubleLucas").
* Files written to are now commited to disk before being closed(Tag: "SLUGFILLER: Commit").
[7] See own credits(Tag: "VQB: ownCredits"):
* Clients on your download queue that have you on their upload queue with a credits mark now have a credits mark.
* You can now see you own UL/DL modifier in a client details, next to the client's UL/DL modifier(In brackets).
[8] New Settings(Tag: "SLUGFILLER: PPgSFVQB"):
* Reconnect on LowID(Tag: "SLUGFILLER: lowIdRetry") - How many retries before accepting a low id(Default: 3).
* Infinite Queue(Tag: "SLUGFILLER: infiniteQueue") - Removes the queue size limit and allows your queue to grow as far as it can. Should be good for the network. Test it out for yourself, and see if there's any down sides(Off by default).
* Multiple Instances(Tag: "VQB: multipleInstance") - You can now choose to allow two or more eMules to run at the same time for whatever reasons you may have(Off by default).
* Slot Shaping(Tag: "VQB: SlotShaping") - Each upload slot gets as much bandwidth as that user can take from the first accepted user in the upload slots to the last, so that user can finish the chunk as fast as possible, while the rest make sure the upload limit is reached. With each user that finishes, the next one in line(according to the order the users got a slot) gets that user's bandwidth. There is a minimal bandwidth per slot, though, which all slots get, even the last one(Off by default). * Disabled pending debug and optimizations *
* Hide overshares(Tag: "SLUGFILLER: hideOS") - Parts in a file uploaded a certain amount of times more than the least uploaded parts are not revealed to other users. This is to prevent a case where all the users download the same part and the other parts become less spread. You can set the amount of times a part has to be uploaded before it's hidden, or 0 to disable(Default: 5).
[9] Detailed downloads:
* xrmb patch for showing in a download's progress bar how much of what was downloaded was already completed and hashed. Slightly modified code(Tag: "xrmb:confirmedDownload").
* xrmb patch for seeing which chunks do sources need from you in the sources status bar(Tag: "xrmb:seeTheNeed").
* Inactive downloads(Paused, error, etc...) have gray-faded status bars(Tag: "SLUGFILLER: grayPause").
* The incomplete side of the progress bar has marks for where chunks start and end(Tag: "SLUGFILLER: chunkDots").
[10] Spread bars v1.2(Tag: "SLUGFILLER: Spreadbars"):
* There is now a "progress bar" in the shared files list showing how many times was each part of a file uploaded.
* Uses black and blue colors, like the download progress bar. Black means never uploaded, light blue means uploaded a little, dark blue means uploaded alot.
* These statistics are saved in the known.met file, using a similar format to the saving of gaps in a part.met file.
* v1.1: You can now sort the files according to the spread bar. Uses a special sorting formula by VQB.
* v1.1: Thanks to VQB 3 new columns were added: the sorting value based on part upload history, a simple total bytes UL/filesize, and the directory the file is in.
* v1.2: Another column, showing how many times a file was shared completely.
[11] Safe Hash(Tag: "SLUGFILLER: SafeHash"):
* The ultimate fix against sharing corrupt downloads, or missing or having corrupt hashlists. See details below.
* Only one known file is hashed at a time. Part file completion is done regardless of currently hashing known files.
* Refreshing the shared files while files are hashing no longer causes these files to be hashed twice.
* Temp files can't be shared, even if the temp directory is shared. Only temp files actually used by eMule are filtered, so other files in the directory can still be shared.
* Closing eMule while files are still hashing no longer causes any crashes, and all threads exit quickly and safely.
* Only complete chunks of part files are rehashed, instead of the entire part file.
* If eMule is closed while a part file is hashing, it will now rehash at startup, instead of assuming it's been hashed.
* All part files hashing, including part finishing and ICH, are now done in another thread.
* Only hashed and verified chunks are shared. Completed chunks are shared only after they are fully hashed. File previewing also checks that the chunks are hashed, and not just complete.
* Loaded hashlists are not assumed to be valid, and are checked against the file hash. Prevents a case of using a corrupt hashlist due to part met file corruption.
* When a hashlist is recieved the part file is instantly hashed with it. Each already completed chunk is shared as soon as they are hashed.
* A part file doesn't wait for the initial hashing to complete before looking for sources and starting to download. It's only shared, though, after it's hashed.
* Only blocks which are currently gaps can be written to, data that has already been downloaded is discarded. This prevents finished blocks from being corrupted by overflowing data.
* Displays number of shared files not yet hashed in statistics, and the expected total number of shared files(including those not yet hashed) in the shared files list title.
Officially Implemented Features:
We stopped implementing these features because the official team either adopted them or came up with their own version. Don't worry, they're still in, the official versions of them, that is.
[1] Adjustable connections per five seconds(fixed, eMule 0.24):
* eMule 0.23 introduced a features which limits how many connections can be made every 5 seconds, so routers aren't flooded due to too many connections in too little time. However, the official client doesn't allow changing the limit. Now you can set whichever limit you choose.
[2] General Bugfixes and Tweaks:
* Shared file list is only flushed to the control itself after eMule has fully initialized. Prevents a crash at startup with Win9X. Based on the patch by Lucas(fixed, eMule 0.25).
* Server.met only read after the Servers tab has fully initialized. Extention of the patch by Lucas(fixed, eMule 0.25).
* Fixed a possible crash if the gap information in a part.met is corrupted, or simply not in a very specific order. Now uses more effecient code for reading an detecting errors in the gap tags(fixed, eMule 0.25).
* Part files without a hashlist are no longer shared as complete sources(fixed, eMule 0.26).
* CPU optimizations by Maella(fixed, eMule 0.26).
* NT/2K/XP users without a C drive couldn't see the directories control(fixed, eMule 0.27).
* When eMule starts the next upload it skips low-id users that aren't currently connected. If a low-id user should have gotten the slot, he is marked, and is automatically given a slot when he next connects(fixed, eMule 0.27).
[3] Safe Hash:
* Gaps are clipped to the file size when loaded. If the part file size is smaller than the expected size then the gap in the end is added in the case it's not already there(Can be caused by crash while flushing the buffer, or by outside modifying if the part file). Gaps are also added in a more controlled manner, which completely prevents gap overlapping(fixed, eMule 0.29).
* Fixed possible bug that only checks the first chunk of each written block for completion, even in the unlikely(but possible) case that two chunks are written to and completed at once(fixed, eMule 0.29).
* Added check when loading the part file to make sure the part file size doesn't exceed the desired size(In case it was enlarged by outside modifying). Prevents bug with both file completing and initial hashing(fixed, eMule 0.29). |