CN1241136C - Method for upgrading software - Google Patents

Abstract

The present invention provides a method for upgrading software. The method comprises the following steps: an increment upgrading packet making process is used for finding out mismatching points of data contents in lower version files and higher version files in order to generate mismatching blocks, corresponding action is generated by aiming at found matching blocks and the mismatching blocks, wherein the action is one of replication, replacement, insertion and deletion; an increment upgrading process is used for executing the action generated in the increment upgrading packet making process in order to replace the contents of the lower version files with the contents of the higher version files. Therefore, the data quantity copied or downloaded in the whole upgrading process is obviously smaller than the data quantity needing to be copied or downloaded by the conventional method, and moreover, the upgrading process is very simple and highly efficient. Therefore, the precious time and the energy of producers and users are saved, and storage medium resources and money are also saved.

Description

Software upgrade method

Technical field

The present invention relates to a kind of software upgrade method.More particularly, the present invention relates to a kind of like this software upgrade method, this method is the software package of highest version with larger software packet upgrade with copy or download low volume data on the lowest version basis of software.

Background technology

Conventional method for upgrading software is to adopt the All Files in the lowest version packet of software, or the mode that a part of file is replaced with whole file is carried out software upgrading.Also do not show the inconvenience that this mode is brought in the software industry early stage of development, because the software in that period mostly is the less packet of data volume, even when each upgrading, each file in the whole packet of redaction is copied one time again, send to the user that need carry out software upgrading then, the time that Software Production merchant's copy and customer upgrade are spent can be not too many yet, and the storage medium of waste is also limited.

But along with software continuous development also, the drawback of conventional method for upgrading software is more and more obvious.On the one hand, the function of software package is more and more, and software size also just correspondingly enlarges.Concerning larger software, conventional method for upgrading software all will copy lot of data in each upgrading, and the AKU that 100,000,000 data occur is much.100,000,000 data will send to the user's hand that carries out software upgrading from the Software Production merchant there, at first, Software Production merchant will spend a large amount of manpower and materials, use jumbo storage medium copy software, the user spends more time that this software upgrade data bag is copied into computing machine more then, to substitute original lowest version software.This process waste many quality time of manufacturer, energy and money.But in fact the content of upgrading accounts for the sub-fraction of the capacity of this packet, each file even may only change several bytes.

On the other hand, some needs frequent upgrading such as the such software package of antivirus software, and the possible data amount of upgrading each time is not very big, but because frequency is very high, in fact causes the waste to user resources yet.If, also wasted the bandwidth of network and the resource of server, increased the weight of the burden of manufacturer by network download.

Therefore be necessary to develop a kind of copy or download the method that necessary minimum contents just can be upgraded software, and needn't be, each all the copy or whole or most of contents of downloaded software for upgrade software.

Summary of the invention

At the problem that exists in the above-mentioned present software upgrading, the purpose of this invention is to provide a kind of more reasonable software upgrade method of more optimizing, this method is on the lowest version basis of software, with copy or the larger software package of download low volume data upgrading.

In order to realize purpose of the present invention, a kind of software upgrade method is provided, this method comprises: increment AKU manufacturing process, be used for finding out lowest version file and highest version file data content do not match a little and behind the match point, do not match piece and a little find out match block to find out with nearest thereafter match point according to match point and nearest thereafter not matching according to not matching a little, wherein produce replication actions at the match block that finds, piece does not produce replacement at matching, insert, one of deletion action, wherein when the piece that do not match does not exist, adopt insert action in the lowest version file; When the piece that do not match does not exist, adopt the deletion action in the highest version file; When the piece that do not match all exists in highest version file and lowest version file, adopt and replace action; With the increment escalation process, be used to carry out the described action that produces by increment AKU manufacturing process, replace to the content of highest version file with content with the lowest version file.This method produces a series of actions by seeking match block.By 4 simple action definitions, as long as the essential data of copy get final product upgrade software to the user.Client is by carrying out increment escalation process upgrade software.Whole upgrade process copy or data downloaded amount are obviously little than required copy of conventional method and data downloaded amount, and escalation process is very simple and efficient, not only save manufacturer and user's quality time and energy, and conserve memory media resource and money.

Description of drawings

The present invention and superiority thereof will be better understood from following detailed description under the situation of reference relevant drawings, wherein:

Fig. 1 shows coupling process flow diagram of the present invention.

Embodiment

Rising Antivirus increment ROMPaq is according to above technical know-how, an embodiment under the WIN32 platform, but the present invention is not limited to the WIN32 platform.

For convenience described below, define several terms now.

The starting point of the identical part of data content is called match point in lowest version file and the highest version file.

The starting point of data content part inequality in lowest version file and the highest version file is called and does not match a little.

Be called match block since a match point (comprising) to a nearest blocks of data district that finishes (not comprising) that do not match.

Be called the piece that do not match since one do not match (comprising) to the blocks of data district that a nearest match point finishes (not comprising).

The preferred embodiment of the present invention realizes software upgrading in two steps: first step is exactly so-called increment AKU manufacturing process, and second step is called the increment escalation process.Describe this two steps now in detail

Increment AKU manufacturing process:

Consider that to the minimum data of user's download carry out operation the most fast, groundwork of the present invention is to finish in making the AKU process, generally this process all is to be finished by manufacturer.

The match point data width of present embodiment is defined as 32 bytes, meets the data width of present main flow Intel CPU.The coupling limit is 4K, meets the joint size of windows platform PE file.

Manufacturing process comprises following step:

1. file is relatively sought the piece that do not match

This is basic step, at first seeks the part that do not match in two files.Compatible portion is that we can repeat to send to user's data, and our whole file comparison procedure is just constantly analyzed the repetition of data and replication problem not in fact.After finding a match block data are labeled as copy.For the piece that do not match, mate again according to step 2.

The identification and matching piece is not that briefly two data equate just passable, because run into through regular meeting in two pieces that do not match that coexist mutually of several bytes yet.So we will ignore too short matched data, this limit of ignoring is also to be to be difficult to hold the very important while.In general, this limit generally should be not less than the data width of data file, and this is usually relevant with operating system.In addition, this limit also will be considered the size of the data structure in the increment AKU, avoids occurring the data block delta package that causes too in small, broken bits and expands.

2. mate the piece that do not match again, according to circumstances produce delta package

The file of two different editions if directly compare, has mass data inequality, that is to say a lot of very big pieces that do not match.This is owing to when program or data modification, often insert and delete content, causes the just dislocation up and down of same data.It is very crucial how reorientating out match point in the piece that do not match, and this is directly connected to the size of the delta package of generation.

The algorithm of matches point basically can reference strings pattern matching algorithm, this algorithm has several, the simplest is that two file interactions are the byte-by-byte comparison of increment with a byte.This method is simple, but can take very many time, and relatively two bigger files of difference may spend several minutes.By improving algorithm, reduce and recall the speed that to increase substantially.

The match point matching algorithm of present embodiment adopts a kind of quick jump searching algorithm, and this algorithm is that the present invention creates, and is adapted at skipping fast in two unmatched data blocks part inequality and seeks match point.At first, we take out the data be used as match point at the lowest version file piece forefront that do not match, and seek first byte of match point then in the piece that do not match of highest version file.If do not find, abandon this match point, move a byte in original, from new coupling.If find first byte, begin to seek second byte of match point from next byte.Do not find with above process, if find each byte of match point fully, the match is successful, if do not wait, recalls a match point length byte number from last matched bytes and restart to take turns matching process.

After finding match point, according to the front and back data cases, the piece that will not match is defined as insertion, deletion or replaces.Be defined as insertion, deletion or replace according to the coupling flow chart description of Fig. 1 piece that how will not match below.

Fig. 1 is the process flow diagram of describing according to a preferred embodiment of the invention.At step S1, carry out pre-service, opening file, to set up reflection, and mate initial point and be set to 0.

At step S2, judge whether to exist the next one piece that do not match.If judge to have data block next inequality in step S2, then this process is transferred to step S3, therein, judges whether not find match point in this matches.

If in this matches piece, do not find match point in step S3 judgement, then this process is transferred to step S4, therein, the match point of new and old edition file is compared, if the match point of highest version file is greater than the match point of ancient deed, then this process is transferred to step S5, and the piece that therein this do not matched is labeled as deleted block, and process is transferred to step S8.

If result in step S4 relatively is the match point of the match point of highest version file less than ancient deed, then this process is transferred to step S6, and the piece that therein this do not matched is labeled as the insertion piece, and process is transferred to step S8.

If result in step S4 relatively is the match point that the match point of highest version file equals the lowest version file, then this process is transferred to step S7, and the piece that therein this do not matched is labeled as replace block, and process is transferred to step S8.

If not can not find match point in step S3 judgement in this matches piece, then this process is transferred to step S7, and the piece that therein this do not matched is labeled as replace block, and process is transferred to step S8.

At step S8, judge that whether this piece that do not match can merge with last one piece that do not match, if draft the result for being, then this process forwards step S9 to, the piece and last that therein this do not matched piece merging that do not match is to produce a bigger piece that do not match.If step S8 judge this do not match piece cannot with on the piece that do not match merge, then this process is transferred to step S10, and this result is write file, and this process is transferred to step S2, to carry out processing procedure next time.

If judge at step S2 not have the next one piece that do not match, then this process forwards S11 to, carries out aftertreatment, with close file.Finish this process then.

The increment escalation process:

As previously mentioned, one of design of uniqueness of the present invention has been with four kinds of simple action descriptions the everything that may do in client, and these four kinds of actions are to duplicate, replace, insert, delete.Duplicating is that data with early version copy to redaction, and this action has a parameter, is exactly length; Replacement is that the data with early version replace to new data, the data length that is replaced in new data length and the early version, and this action has two parameters, and the one, the length of replacing, the 2nd, new data; Insertion is that new data are inserted in the data of early version, forms the data of redaction, and this action has two parameters, the one, and the length of insertion data, the 2nd, the data of inserting; Deletion is the data deletion with early version, because there are not these data in redaction.Use above-mentioned four kinds of actions, can finish the upgrading to single file, process is as follows:

(1), open early version file to be upgraded, whether the content of checking early version is the version that will upgrade, if not, can't upgrade, withdraw from escalation process; If create the redaction file.

(2), get the action and parameter, for different actions, do following processing respectively:

1, duplicate, parameter is for duplicating length.From the early version file, fetch data, the data that read are write new file by duplicating length read.

2, replace, parameter is for replacing length, replacement data.The file pointer of early version file is moved backward, and displacement is for replacing length; Replacement data is write the redaction file, write length for replacing length.

3, insert, parameter is for inserting length, inserting data.To insert data and write the redaction file, write length for inserting length.

4, deletion, parameter is deletion length.The file pointer of early version file moves backward, and displacement is for replacing length.

(3), repeated for second step, to finishing to the everything of this document.

(4), close the early version file, close the redaction file, finish upgrading to a file.

(5), repeat first to fourth the step, finish escalation process to all files to be upgraded.

The benefit that this upgrade method is brought is:

1. produce in the process of highest version file, no matter lowest version file or highest version file, file refers to Pin is all without any trace-back process, so that file read-write amount minimum, and speed piece very not only, and really do Have no requirement to the configuring condition for the user.

2. four simple actions can be contained all complicated file update operations, so that the targeted update mistake Journey is isolated fully with the operating system and the algorithm that produce AKU. Manufacturer at all can shadow to the adjustment of algorithm Ring user's upgrading. For the user of different operating system, manufacturer also needn't make different AKUs.

3. make client work very simple, core code only needs several row just can finish all working.Client-side program is very short and small and stable.

Claims (4)

1. method for upgrading software, this method comprises:

Increment AKU manufacturing process, be used for finding out not matching a little and behind the match point, not matching piece and a little find out match block of lowest version file and highest version file data content according to match point and nearest thereafter not matching to find out with nearest thereafter match point according to not matching a little;

Wherein produce replication actions at the match block that finds;

Produce replacement, insert, delete one of action at the piece that do not match,

Wherein when the piece that do not match does not exist, adopt insert action in the lowest version file; When the piece that do not match does not exist, adopt the deletion action in the highest version file; When the piece that do not match all exists in highest version file and lowest version file, adopt and replace action; With

The increment escalation process is used to carry out the described action that is produced by increment AKU manufacturing process, replaces to the content of highest version file with the content with the lowest version file.

2. according to the method for upgrading software of claim 1, the data width of wherein said match point is 32 bytes of fixing.

3. according to the method for upgrading software of claim 1, wherein this method also is included in the described piece that do not match and seeks match point again,

Wherein do not match and seek the ultimate range of match point in the piece again, promptly mating limit is 4K.

4. according to the method for upgrading software of claim 1, the matching algorithm of wherein said match point is quick jump searching algorithm.